Integrative Approaches in Environmental Health and Exposome Research: Epistemological and Practical Issues 3031284313, 9783031284311

Citation preview

Integrative Approaches in Environmental Health and Exposome Research Epistemological and Practical Issues Edited by Élodie Giroux · Francesca Merlin · Yohan Fayet

Integrative Approaches in Environmental Health and Exposome Research

Élodie Giroux · Francesca Merlin · Yohan Fayet Editors

Integrative Approaches in Environmental Health and Exposome Research Epistemological and Practical Issues

Editors Élodie Giroux Philosophy Department Jean Moulin University Lyon 3 and Institut de Recherches Philosophiques de Lyon Lyon, France

Francesca Merlin Institut d’histoire et de philosophie des sciences et des techniques (IHPST) CNRS and University of Paris 1 Panthéon-Sorbonne Paris, France

Yohan Fayet Human and Social Sciences Department, Centre Léon Bérard Lyon, France UMR Inserm U1290 RESHAPE Lyon, France UMR 5600 Environnement, Ville, Société Lyon, France

ISBN 978-3-031-28431-1 ISBN 978-3-031-28432-8 (eBook) https://doi.org/10.1007/978-3-031-28432-8 © The Editor(s) (if applicable) and The Author(s) 2023 This work is subject to copyright. All rights are solely and exclusively licensed by the Publisher, whether the whole or part of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission or information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now known or hereafter developed. The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant protective laws and regulations and therefore free for general use. The publisher, the authors, and the editors are safe to assume that the advice and information in this book are believed to be true and accurate at the date of publication. Neither the publisher nor the authors or the editors give a warranty, expressed or implied, with respect to the material contained herein or for any errors or omissions that may have been made. The publisher remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Cover illustration: Sue Chillingworth/Alamy Stock Photo This Palgrave Macmillan imprint is published by the registered company Springer Nature Switzerland AG The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland

Acknowledgements

This volume follows on from a two-day international meeting led in Lyon on February 4–5 2021, organized by the EPIEXPO group (“Critical epistemology of the exposome”) directed by Élodie Giroux. This meeting and the present volume were supported by the Convergence PLascan Institute (ANR-17-CONV-0002 PLASCAN) in Lyon and by the ANR EnviroBioSoc Project (ANR-19-CES26-0018-01, directed by Francesca Merlin). The editors warmly thank the authors for their time and commitment that made this volume possible. We also thank the anonymous reviewers for their insightful comments which allowed to improve the quality of this volume.

v

Contents

Airs, Waters, Places… and the Exposome: Steps Toward an Integrative Health Maurizio Meloni

1

Articulating the Social and the Biological The Turn Towards ‘The Biosocial’ in Epigenetics: Ontological, Epistemic and Socio-Political Considerations Luca Chiapperino Socio-Markers and Information Transmission Federica Russo What’s Wrong with the Biologization of Social Inequalities in Health? A History of Social Epidemiology and Its Moral Economy of Objectivity Mathieu Arminjon

9 35

65

Integration in Environmental Health and Exposome Research: Epistemological Issues Which Integration for Health? Comparing Integrative Approaches for Epidemiology Stefano Canali

101

vii

viii

CONTENTS

A Critical Assessment of Exposures Integration in Exposome Research Élodie Giroux From Exposome to Pathogenic Niche. Looking for an Operational Account of the Environment in Health Studies Gaëlle Pontarotti and Francesca Merlin

129

173

The Case of Exposome Research: Practical and Disciplinary Issues Place of Integrative Approaches in the Study of Spatial Dimension of Health Outcomes Yohan Fayet

209

The Exposome and the Social Sciences: The Case of Systemic Diseases Catherine Cavalin

239

The Exposome Research Program and Nutrition: The Example of Celiac Disease Paolo Vineis and Antonio Francavilla

259

Index

269

Notes on Contributors

Mathieu Arminjon is a philosopher and historian of biomedical sciences and care. He works at the School of Health Sciences (HESAV), University of Applied Sciences and Arts Western Switzerland (HES-SO), in Lausanne. He has written various articles and books on the history of neuroscience, the physiology of stress and social epidemiology. He is currently working on a project funded by the Swiss National Science Foundation aiming at studying the development of research on the social determinants of health from an agnotological perspective (SNSF 200460). Stefano Canali works in philosophy of science, with particular interests in medicine, data and technology. He’s currently postdoc in the Department of Electronics, Information and Bioengineering of the Politecnico di Milano, with a project on the epistemology and ethics of wearable technology for health. He did his Ph.D. at Leibniz Universität Hannover on the epistemology of the exposome and data-intensive science. Previously, he studied philosophy of science and science and technology studies at University College London and philosophy at the University of Milan. Catherine Cavalin is a permanent fellow researcher in sociology at IRISSO (Interdisciplinary Research Institute in the Social Sciences), National Centre for Scientific Research (CNRS). She works on the diversity of health statuses and social health inequalities, which include gender, labour and exposure to toxicants at work, as well as interpersonal

ix

x

NOTES ON CONTRIBUTORS

violence. Her research systematically encompasses a historical and sociological approach of knowledge. She particularly investigates the categories on which statistics are based and the nosological categories that frame medical knowledge. In 2021, she co-edited Cent ans de sous-reconnaissance des maladies professionnelles (with Emmanuel Henry, Jean-Noël Jouzel and Jérôme Pélisse; Paris, Presses des Mines), and in 2022, co-edited Les violences sexistes après #MeToo (with Jaércio da Silva, Pauline Delage, Irène Despontin Lefèvre, Delphine Lacombe and Bibia Pavard; Paris, Presses des Mines). Luca Chiapperino is Swiss National Science Foundation (SNSF) Ambizione Lecturer at the STS Lab, Faculty of Social and Political Sciences (SSP), University of Lausanne (Switzerland). Following his Ph.D. in “Foundations of the Life Sciences and their Ethical Consequences” at the European School of Molecular Medicine (SEMM) and the University of Milan (main advisor: Prof. Giuseppe Testa), he coled several SNSF projects with professors Francesco Panese (UNIL) and Umberto Simeoni (CHUV). He is currently the Principal Investigator of the Ambizione project “Constructing the Biosocial: an engaged inquiry into epigenetics and post-genomic biosciences ” (Grant N. 185822). His research interests are situated at the crossroads of Science and Technology Studies (STS) and applied philosophy, a position from which he studies the mutual shaping of the epistemic and socio-political dimensions of biomedical research. His work has appeared in numerous specialised journals, book collections, as well as in interdisciplinary publications. Yohan Fayet holds a Ph.D. in Geography (Lyon 3 University, 2014) and is research project manager in Human and Social Sciences Department at the Léon Bérard Center in Lyon. He is interested in the analysis of spatial inequalities in health, through the multiple impact of physical, social and medical environments on health outcomes. He focuses on spatial inequalities in cancer and analyses potential challenges raised by innovations in oncology. He published a nationwide geographical classification in France to measure spatial inequalities in health (International Journal of Health Geographics, 2021) and works on potential actions to reduce it.

NOTES ON CONTRIBUTORS

xi

Antonio Francavilla is a molecular biologist at the Italian Institute for Genomic Medicine of Turin, Italy. His research interest concerns molecular and genetic epidemiology in the context of gastrointestinal disorders and nutrition. He is co-author of relevant articles on these topics. Lately, he has mainly been involved in the research of non-invasive molecular biomarkers of diagnosis and monitoring of Celiac Disease, a project found by the Associazione Italiana Celiachia (AIC). Élodie Giroux is philosopher of science, medicine and epidemiology. She is Professor at Lyon 3 University and Researcher at the Lyon Institute of Philosophical Researches. Her main publications are on historical epistemology of risk factor epidemiology, philosophy of epidemiology, causation in medicine and public health, health and disease concepts, and more recently, precision medicine and integrative research in environmental health and exposomics. Besides several papers on those topics, she published Après Canguilhem, définir la santé et la maladie (2010), and she edited Naturalism in the philosophy of health (2016) and several special issues. She coordinates the EPIEXPO (‘EPIstemology of EXPOsome’) project in the context of Plascan Institute in Lyon. Maurizio Meloni is a social theorist and a science and technology studies scholar. He is the author of L’Orecchio di Freud. Societa’ della comunicazione e Pensiero Affettivo (Dedalo, 2005); Political Biology: Science and Social Values in Human Heredity from Eugenics to Epigenetics (Palgrave 2016); Impressionable Biologies: From the Archaeology of Plasticity to the Sociology of Epigenetics (Routledge, 2019); co-editor of Biosocial Matters (Wiley 2016); and chief editor of the Palgrave Handbook of Biology and Society (2018). He is currently an ARC Future Fellow and Associate Professor of Sociology at Deakin University, Australia. He has benefited from several research grants, including two Marie Curie fellowships, a Fulbright scholarship, funded visits at the Max Planck Institute for the History of Science (MPIWG, Berlin), DAAD and OEAD fellowships in Germany and Austria, and an annual membership at the Institute for Advanced Study in Princeton (NJ). Francesca Merlin, permanent research fellow in philosophy of science at CNRS (IHPST lab), holds a Ph.D. in Philosophy (University of Paris 1 Panthéon Sorbonne, 2009). Her research focuses on central concepts in biology such as chance and probability, inheritance and epigenetics, in particular in the context of evolutionary theory. She published several

xii

NOTES ON CONTRIBUTORS

papers on these topics and the book Mutations et aléas : le hasard dans la théorie de l’évolution (Hermann, 2013). She also co-edited, with Thierry Hoquet, Précis de philosophie de la biologie (Vuibert, 2014). She currently coordinates the EnviroBioSoc project, an interdisciplinary research project funded by ANR and dealing with the plurality of ways in which the environment is conceived and operationalised in the study of environmentally induced diseases throughout biomedical and social sciences. Since 2018, she is President of the Société de Philosophie des Sciences (SPS). In 2019, she was awarded by the CNRS Bronze Medal. For more information, visit her personal website: https://sites.google. com/site/francescamerlin/. Gaëlle Pontarotti holds a Ph.D. in Philosophy from University of Paris 1 Panthéon Sorbonne (2017). She is a philosopher of biology and a philosophy teacher. Her research deals with the concepts of extended heredity, epigenetics, environment and race. She has written various articles on these topics. She has also co-edited a volume dedicated to the concept of Identity: L’identité, Dictionnaire encyclopédique (Gallimard, 2020). She was a postdoctoral fellow in the project EnviroBioSoc, funded by ANR. Federica Russo is a philosopher of science, technology and information. She is Honorary Professor at University College London (Department of Science and Technology Studies) and lectures at the University of Amsterdam (Department of Philosophy and Institute for Interdisciplinary Studies). At the UvA, she carries out research at the Institute for Logic, Language and Computation, and within the Language and Cognition in Argumentation Group. Her research concerns epistemological, methodological and normative aspects they arise in the health and social sciences, with special attention to policy contexts and to the highly technologised character of these fields. She has published extensively on various themes, such as causation and causal modelling, evidence and technology, and her latest monograph is titled Techno-Scientific Practices : An Informational Approach (RLI, 2022). She has been co-editor in chief (with Phyllis Illari) of the European Journal for Philosophy of Science and is currently Editor-in-Chief of Digital Society. She sits in the Management Team of the Institute for Advanced Study at the University of Amsterdam and is member of the Steering Committee of the European Philosophy of Science Association. For more information please visit russofederica.wordpress.com or follow her on Twitter @federicarusso.

NOTES ON CONTRIBUTORS

xiii

Paolo Vineis is Chair of Environmental Epidemiology at Imperial College London. He is a leading researcher in the field of molecular epidemiology. His latest research focuses on environmental exposures and intermediate markers from omic platforms in large epidemiological studies. He also investigates the effects of climate change on noncommunicable diseases. He is a principal investigator or co-investigator of numerous international projects.

List of Figures

Socio-Markers and Information Transmission Fig. 1

Relations between bio-markers and causes

37

The Exposome Research Program and Nutrition: The Example of Celiac Disease Fig. 1

In high-income countries (HIC), individuals with low socio-economic status (SES) are presented with several significant complications that may both directly and indirectly influence the composition of the microbiota and, concurrently, the immune system. Specifically, we suggest that environmental stressors, the pressure to eat affordable food that is filling and palatable, along with alterations in health care and medication use, can hamper microbiota diversity and promote a low-grade inflammatory state that precipitates metabolic disease. NCD, non-communicable disease; T2D, type 2 diabetes (reproduced from Harrison and Taren 2018)

265

xv

List of Tables

Which Integration for Health? Comparing Integrative Approaches for Epidemiology Table 1

Approaches to data integration analysed with respect to their specific features and limitations

119

The Exposome Research Program and Nutrition: The Example of Celiac Disease Table 1

Classification of autoimmune diseases

262

xvii

Airs, Waters, Places… and the Exposome: Steps Toward an Integrative Health Maurizio Meloni

A decade ago, prominent US medical historian Charles E. Rosenberg (2012) wrote about the surprising return of an “epidemiology of place” that has ultimately its roots in the Hippocratic tradition of Airs, Waters, and Places (5th c BCE). Increasingly marginalized during the twentieth century, though never displaced (Anderson 2004), an emphasis on places and the socio-physical environment receded on the background vis-a-vis the rise of the abstracted body of bacteriology, biochemistry, and later genomics. However, catalyzed by the emerging awareness of a global ecological crisis and its health implications, the inescapable entanglement of bodies and places when it comes to health has captured again the center scene. In Rosenberg’s words, “in our reengineered—post-Darwin and post–molecular biology” epoch, “we have recreated our particular understanding of airs, waters, and places, of how the body functions in the world and the world shapes our bodies” (2012: 668; see also for a wider

M. Meloni (B) Faculty of Art and Education, Alfred Deakin Institute, Deakin University, Melbourne, VIC, Australia e-mail: [email protected]

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 É. Giroux et al. (eds.), Integrative Approaches in Environmental Health and Exposome Research, https://doi.org/10.1007/978-3-031-28432-8_1

1

2

M. MELONI

context, Bashford and Tracy 2012). Of course, this is not to draw shallow analogies or flat continuities between, say, the exposome diagram drawn by Wild in 2012 and centuries of medical diagrams where the Hippocratic body is at the center of a range of influences, such as climate, winds, seasons, age, and stars. Undoubtedly, there are obvious differences between the mostly man-made factors of exposures in the exposome framework and the relatively natural forces of climate in the Hippocratic tradition. Another major difference is that science before the rise of mechanical philosophy in the seventeenth century didn’t aim to quantify “exposures” in the way we do today.1 However, interesting parallelism or resonances cannot be denied either. Firstly, following Rosenberg’s analysis, an emphasis on the “body as always situated and always in process,” and “biological individuality (as) cumulative and disease (as a) multicausal and multidimensional phenomenon” (2012). Secondly, today as in the Hippocratic tradition, the “environment” “is an aggregative combination of various elements which interfere with each other to produce certain damage or benefits on populations and individuals” (Le Blay 2005, 253). Thirdly, one could add, an interest in the porous boundaries between what is external and what internal to the body, exactly that transition zone where one turns into the other: “the internal component of exposures” in today’s language; how “toxic materials become lodged in all the fatty tissues of the body” in Rachel Carson’s prophetic analysis (1962); or, in Ibn S¯ın¯a’s Al-Qanun fi’l-tibb or Canon of Medicine (c. 1025), the most influential medical handbook in European Universities until the fifteenth century, “septic changes in the body fluids [humours]” caused by corrupted air. All these statements, across different epochs and contexts, trouble the certainty of a neat separation between the external and the internal to the body or, to over-simplify, the disentangled and abstract body ontology of modern biomedicine. I am suggesting this longue-durée context to the present volume on biosociality, integrative health, epidemiology, and the exposome not in a spirit of provocation. To me, the exposome-to-Hippocrates nexus (via the

1 The two statements above are not meant to reinforce the stereotype that ancients did

not understand ecological crisis as effects of social factors including deforestation, military occupation, exploitation of lands, because they obviously did (see Harper 2017). It is also not to reinforce the belief that medicine before modernity was devoid of a subtle logic of calculation. It was in fact very invested in notions of ratio, proportion, and balance (Kaye 2014).

AIRS, WATERS, PLACES… AND THE EXPOSOME: STEPS …

3

environmental movement) evidences a problem that lies behind many of the chapters of this important book by emerging and senior European researchers. I would summarize the point by saying that all chapters point to a certain dissatisfaction with the received notion of the environment and the series of binaries it produces: that is, binaries between organism and their surroundings, the biological (“inside the skin”) and the social (“outside the skin”), agency and passivity of the organism. A notion of relatively recent coinage, the modernistic term environment, understood as a universal medium surrounding bodies, seems to assume an original separation of bodies from their external circumstances to look for a subsequent connection or integration at a later stage. If we follow Canguilhem’s story (2001) about the genealogy of the term milieu, from which the British environment reappeared in English in 1853 (Pearce 2010), one of its key roots is in the mechanistic philosophy of the eighteenth century. In d’Alembert and Diderot’s Encyclopédie (1751), we can read for instance that milieu “means a material space through which a body passes through its movement, or in general, a material space in which a body is placed, whether it is moving or not. Hence, we can imagine ether as a milieu in which celestial bodies (corps célestes ) move” (my translation, emphasis added). Canguilhem correctly points out that the story of its migration to biology via Buffon and Lamarck, and later Comte, is complex, and other influences on the term milieu are evoked, including the Hippocratic Airs Waters and Places. However, if the meaning fixed by the French Encyclopedists determines the later usage of the term, and its great scientific appeal, we can start to see why this notion of a medium surrounding bodies or through which bodies circulate does not capture the intensity of the embodiment of exposures we find in the Anthropocene and its toxic legacy (Lock 2018). Nor, to capture the current porosity and plasticity of the Earth and biological systems to human action (Meloni et al. 2022), seems enough to recover its original meaning of medium “between two centers,” as Canguilhem suggested in the 1960s. Even allowing for very little mechanicism and disenchantment in Newton’s notion of action at a distance (Storm 2017), the environment as background conditions and a network of external stimuli does not capture well the precariousness of the body/environments boundary that we found both in the Hippocratic tradition (and more widely premodern medicine and indigenous ontologies) and in contemporary biosocial and postgenomic frameworks. I am not saying that the modernistic term “environment” is imploding under the heavy toxicity of

4

M. MELONI

the Anthropocene. However, its consistence and capacity to explain ubiquitous phenomena of toxic exposures and “exposed biologies” (Wahlberg, 2018), diversified effects of life experiences and social conditions (from food to systemic racism and in general “slow violence”), and dynamic coconstruction of the milieu through organismic influences, are increasingly challenged (Landecker, 2011; Niewöhner and Lock, 2018). The instability of the modernistic division between bodies and their medium, as well as the need for a more complex, biosocial, and ecological understanding of organism/environment interaction, is what from different disciplinary backgrounds the different chapters in the book insightfully address. They look at incommensurability and different possibilities to bridge the social/biological divide (Chiapperino, Russo), at the complex and unstable politics of biologization (Arminjon), at new models of integrative health and exposure science and their often unknown family tree in the history of social epidemiology (Canali, Giroux, Vineis and Francavilla, Cavallin, and Fayet), at the inadequacy of the environment as conceived in exposome research and the need of a better philosophy of biology to frame science (Pontarotti and Merlin). Overall, these chapters show epistemological tensions in the construction of postgenomic sciences and the search for a more dynamic way of thinking the nexus of bodies and environments. As such, they may be taken as evidence for some sort of post-paradigmatic transition that will possibly replace “the environment” with something more granular, dialectic, porous, and capable to capture the plasticity of both Earth and biological systems across multiple scales of analysis. Acknowledgments Maurizio Meloni gratefully acknowledges support by an Australian Research Council Future Fellowship (FT180100240).

References Anderson, W. (2004). Natural histories of infectious disease: Ecological vision in twentieth-century biomedical science. Osiris, 19, 39–61. Bashford, A., & Tracy, S. W. (2012). Introduction: Modern airs, waters, and places. Bulletin of the History of Medicine, 86(4), 495–514. Carson, R. (2002 [1962]). Silent spring. New York: Houghton Mifflin Harcourt. Canguilhem, G. (2001). The living and its milieu. Grey room, 3, 7–31. Harper, K. (2017). The fate of Rome. Climate, disease and the end of an empire. Princeton University Press.

AIRS, WATERS, PLACES… AND THE EXPOSOME: STEPS …

5

Kaye, J. (2014). A history of balance, 1250–1375. Cambridge: Cambridge University Press. Landecker, H. (2011). Food as exposure: Nutritional epigenetics and the new metabolism. BioSocieties, 6, 167–194. Le Blay, F. (2005). Microcosm and macrocosm: The dual direction of analogy in Hippocratic thought and the meteorological tradition. In P.J. van der Eijk (ed.) Hippocrates in context. Brill, p. 251–269. Lock, M. (2018). Mutable environments and permeable human bodies. Journal of the Royal Anthropological Institute, 24(3), 449–474. Meloni, M., Wakefield-Rann, R., & Mansfield, B. (2022). Bodies of the Anthropocene: On the interactive plasticity of earth systems and biological organisms. The Anthropocene Review, 9(3), 473–493. Niewöhner, J., & Lock, M. (2018). Situating local biologies: Anthropological perspectives on environment/human entanglements. BioSocieties, 13, 681–697. Pearce, T. (2010). From ‘circumstances’ to ‘environment’: Herbert Spencer and the origins of the idea of organism—Environment interaction. Studies in History and Philosophy of Science Part C: Studies in History and Philosophy of Biological and Biomedical Sciences, 41(3), 241–252. Rosenberg, C. E. (2012). Epilogue: Airs, waters, places. A status report. Bulletin of the History of Medicine, 86(4), 661–670. Storm, J. A. J. (2017). The myth of disenchantment: Magic, modernity, and the birth of the human sciences. University of Chicago Press. Wahlberg, A. (2018). Exposed biologies. Good quality: The routinization of sperm banking in China. Oakland, CA: University of California Press.

Articulating the Social and the Biological

The Turn Towards ‘The Biosocial’ in Epigenetics: Ontological, Epistemic and Socio-Political Considerations Luca Chiapperino

1

Introduction

Research in epigenetics affirms—or rather revives1 —the centrality of plasticity thinking and the importance of (ecological, social, cultural) environments in the understanding of health and disease. Yet, the scope, extension and implications of this generic claim are to be fully spelt out. As argued by (yet another) recent authoritative ‘critical’ appraisal

1 Against historical simplifications of epigenetics and its role in the shift towards postgenomic explanations of biological phenomena, several scholars have shown how the plasticity thinking in development and evolution populating contemporary epigenetics runs throughout the history of biology (Pigliucci 2001; Jablonka and Lamb 2005; Meloni 2016; Morange 2018; Chiapperino and Panese 2019).

Funding: This work is part of the Swiss National Science Foundation (SNSF) Ambizione project “Constructing the Biosocial: an engaged inquiry into epigenetics and post-genomic biosciences” (PZ00P1_185822). L. Chiapperino (B) STS Lab, Faculty of Social and Political Sciences, Institute of Social Sciences, © The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 É. Giroux et al. (eds.), Integrative Approaches in Environmental Health and Exposome Research, https://doi.org/10.1007/978-3-031-28432-8_2

9

10

L. CHIAPPERINO

of epigenetics, definitional ambiguities persist around the term epigenetics (Horsthemke 2022) in ways that mimic “a pronounced dichotomy within the field” (Deans and Maggert 2015, 889). On the one hand, some employ epigenetics to designate the study of phenotypic plasticity. This version of epigenetics—which emerged in the path of Conrad Hal Waddington in the 1940s (Waddington 2012)—chases the biological mechanisms of genotype to phenotype transitions and the ways these processes are moulded by the environment and developmental trajectories. Taking full stock of this Waddingtonian thread of research, some have argued, suggests studying biology’s intricate dependence on developmental conditions and environmental stimuli, much like abandoning dichotomies between nature and nurture in the understanding of development, health and evolution (Jablonka and Lamb 2005; Fox Keller 2010; Lock 2015). On the other hand, others take epigenetic research to consist of studies of gene expression and how this programming persists across different cell generations (i.e. mitosis) or—but only according to some— across organismic generations (i.e. meiosis). This molecular approach to the study of epigenetic regulation of gene expression emerged in the mid1950s (Nanney 1958) and retains a strong hold over epigenetics today. Epigenetics is, in fact, often defined as the study of the molecular modifications of DNA that change its expression without altering its sequence and are inherited through mitotic/meiotic processes (Bird 2007). There exist open questions on both sides of this dichotomy. Even further, there exist also a wealth of positions that, with different degrees of clarity and ambiguity, sit in between these two extremes (Stotz and Griffiths 2016). It suffices to say here that epigenetics is an open front and that these different programmes touch, in different ways, upon foundational notions in biology such as ‘genes’, ‘genomes’ and ‘organisms’ (Fox Keller 2015; Rheinberger and Müller-Wille 2018). Even the strong molecular programme in epigenetics—which is the most contiguous to genetic reductionism (Stotz and Griffiths 2016)—breaks with genecentrism: it points to the need of understanding how much individual gene expression differences (e.g. in transcription factors, proteins or genomic regulatory regions) is actually the product of environmental stimuli and individual epigenetic differences (Jaenisch and Bird 2003;

University of Lausanne, Lausanne, Switzerland e-mail: [email protected]

THE TURN TOWARDS ‘THE BIOSOCIAL’ IN EPIGENETICS …

11

Cavalli and Heard 2019). Epigenetic research concerned with developmental and environmental factors raises instead questions about conceptions of the body and the genome in contemporary life sciences: rather than being allegedly fixed entities (regulated by a genetic programme crafted by evolutionary temporalities and processes) the genome and the body in (this strand of) epigenetics are porous, malleable, plastic entities resulting from unique processes of mixed biological, social, environmental, developmental differentiation (Lickliter and Witherington 2017). While reliant on the molecular biology tools that follow the Human Genome Project, epigenetics—in its different and idiosyncratic versions— is a post-genomic research programme which emphasises in many different ways “complexity, indeterminacy and gene-environment interactions” in the study of biological processes (Richardson and Stevens 2015, 3). It is thus no surprise that epigenetics raises a great deal of interest in the social sciences and humanities. Many scholars in Science and Technology Studies (STS), philosophy of biology and applied ethics increasingly recognise the opportunity epigenetics offers to install a shared theoretical and integrative approach to human health in our societies (Rose 2013; Niewöhner 2015). As several have asked: can this systematic effort to disentangle the biological and environmental modulators of health and disease be an integrative and interdisciplinary science of health promotion? Specifically, the opportunity epigenetics offers to study the biology of health differentiation in the face of social conditions is often qualified with the adjective ‘biosocial’ (Meloni et al. 2018a). There exist biosocial research agendas crystallising around epigenetics (Dubois et al. 2020); epigenetics is a prominent element of the biosocial landscape of worldviews highlighting the importance of social-biological transitions in health (Meloni et al. 2018a); epigenetics also instructs strategies of intervention in these processes that can be qualified as biosocial because they activate body-environment porosity in dealing with molecular predispositions to disease (Chiapperino 2021). Biosocial occurs also as noun in the form of ‘biosociality’—a term that has been used to criticise the political implications of an increasingly blurred distinction between the social and the biological since the 1990s (Meloni et al. 2018a, 19; cf. Rabinow 1996). STS and humanities scholars have highlighted several critical aspects in the life sciences’ production of factual and socio-political resources to understand biosocial representations of health in epigenetics (Meloni et al. 2018a). ‘The biosocial’ is pinned against the reductionist and determinist

12

L. CHIAPPERINO

thinking that still permeates the explanatory frameworks of biology (PittsTaylor 2019)—even when the life sciences specifically problematise the biosocial entanglements that produce organismic development and health differentiation. This chapter identifies the challenges and gaps that keep the qualifier ‘biosocial’ away from the methods, facts and policy translations of epigenetic research. More than to systematise the concept in a specific definition, my objective here is to flesh out the incommensurability of the biosocial styles of reasoning across the social and life sciences (Fleck 1981; Hacking 1992; Sciortino 2017; Sady 2021). Juxtaposing the ways STSers characterise ‘the biosocial’ and the biosocial openings in epigenetics could establish the nuances, gaps and practical changes that make it currently impossible to translate ‘the biosocial’ from social sciences and humanistic thought styles into the one of the life sciences (and viceversa). While this issue is one of communicative interactions and language (as Fleck’s notion of thought collective would suggest), I argue that incommensurability here is fundamentally tied to the research methods and standards of evidence of different styles of reasoning (thus echoing Hacking’s epistemological reflections on these matters). To this purpose, the chapter delineates the ontological, epistemic and sociopolitical dimensions of ‘the biosocial’. The incommensurability between epigenetic studies of social-biological transitions in health and humanistic conceptions of health as entangled, hybrid and biosocial comes in different versions. These include the difficulty of converging into a shared health research (Müller et al. 2017), but also specific conceptions of the body as biosocial entity (Niewöhner and Lock 2018). Another gap relates to the idea, which is largely missing from epigenetic research, that health promotion policies ought to necessarily adopt a syndemic perspective attentive to the mixed social and biological determinants of health (Singer et al. 2017). By singling out these different dimensions of biosocial ideas and research practices, my hope is to improve the opportunities for collaboration around the entangled, hybrid and situated nature of health. The turn towards ‘the biosocial’ in epigenetics, I conclude, presents practitioners—much like social and humanistic critics—with the necessity to situate such questions of incommensurability and to give way to a stance privileging a focus on socio-political change in interdisciplinary biosocial experimentations.

THE TURN TOWARDS ‘THE BIOSOCIAL’ IN EPIGENETICS …

2

13

Biosocial Ontologies of the Living

At a first, and more fundamental level, constructing the biosociality of health and disease in epigenetics entails a specific type of ontological work. This includes enacting conceptions of the continuity between bodies and their environments, offering assumptions as to how to make sense of these processes, as well as producing entities (e.g. markers, mediators, mechanisms) that account for these entanglements in knowledge practices. These ontological ramifications of the concept find their first clear-cut formulation in the work of anthropologist Ashley Montagu (Montagu 1956). His book “The Biosocial Nature of Man” takes a genealogical approach to the philosophical question of human nature, to provide a historical dissolution of dichotomous conflictual understandings of humans as “animal[s] and as […] social creature[s]” (p. 9). In Montagu’s view, human constitution is a process, which depends as much from “what has been acquired by the organism from its environment” as from “what it has acquired from its genetic endowment” (p. 41). Recognising the relevance of these biosocial processes is not just a philosophical issue; it is also a fundamental step to change social organisation and operate around notions of race, gender and criminality (to name his own examples). Highlighting the biosocial constitution (or ontology, I shall say) of humans has in fact the consequence of putting at the centre of social development the importance of “education” more than “predestination” (pp. 108–109). As Montagu argues, human nature “is learned and acquired within limits of those uniquely human potentialities for being human in a particular culture” (p. 80). Within recent social sciences scholarship, these ideas have made a comeback due (among other things) to the opportunities opened by research fields such as epigenetics. The question that has preoccupied much STS and philosophical scholarship is whether glimpses of such a biosocial humanism can be found in these experimental practices. The ontology of the biosocial STS scholars have in mind is a material-semiotic process situated in evolutionary, intergenerational, developmental, lifecourse, metabolic times, as much as in cellular, organismic, molar, social spaces. It is also an ontology of the biosocial that belongs to patterns of practice that are experimental, linguistic, symbolic, medical, cultural and cross-cultural (Niewöhner and Lock 2018). Others connect biosocial ideas to a processual ontology of life (Nicholson and Dupré 2018) in ways that find both correspondences with strands of complexity

14

L. CHIAPPERINO

thinking in the philosophy of biology (Stotz and Griffiths 2016), all while tying into the organicist tradition at the roots of modern epigenetics (Peterson 2017). Biosocial ontologies of this kind also connect epigenetic evidence to strands of feminist thinking, as several have pointed out (Meloni et al. 2016; Meloni et al. 2018a; Niewöhner and Lock 2018). Tracing these ideas back to Donna Haraway (Haraway 1990) and Karen Barad (Barad 1998), these authors have emphasised in different ways how epigenetics raises, from within the life sciences, the question about the co-constitution of life as both a material and a semiotic-discursive phenomenon. To paraphrase Margaret Lock, many of the findings of epigenetic research are “presumed to substantiate a mechanism whereby nature and nurture meld as one” (Lock 2020, 26). And yet, the author asks, is the idea of life as a situated process in the making really contiguous to the knowledge base of epigenetics? It is a matter of contestation whether epigenetics embraces such an ontology of life as ‘becoming’ (Ingold and Pálsson 2013; Niewöhner and Lock 2018). In fact, the importance critics put on the recognition of a symmetrical and processual ontology of biological and social processes in epigenetics hardly finds correspondence in what has been observed in scientists’ work. Epigenetics’ reunification of the biological and the social, critics argue, looks more like the miniaturisation of these processes (Lock 2013, 2020) than a study of health “differentiation as a material-semiotic practice” (Niewöhner 2020, 57). Epigenetic scientists often study the social environment as a variable to be captured by proxy measures such as socio-economic status (Evans et al. 2021). The relevance of these proxy measures gets in turn interpreted on the basis of correlational measures of biological differences, such as DNA methylation (Cerutti et al. 2021). Of note, measures of social position are seldom standardised and lack any theorisation or uptake of the multiple causalities and looping effects between biological and social determinants of health (Darling et al. 2016; Evans et al. 2021). In a nutshell, “nature and culture” get construed in epigenetics as “additive” (Niewöhner 2020, 55): an element that gives a foundation to these biosocial processes in epigenetics which is thoroughly different from the one circulating in STS and humanities circles. Ontologies mediating for the traffic between social experiences and functional modifications of the genome in epigenetics often consist of ‘molecular conduits’ for body-environment traffic (Landecker and Panofsky 2013) more than the affirmation of a holistic systems view of the organism (Oyama 2000).

THE TURN TOWARDS ‘THE BIOSOCIAL’ IN EPIGENETICS …

15

Thus, practices of epigenetic research enact specific realities of biosocial processes, which are not devoid of (ontological) consequences. Epigenetic research crafts specific patterns of relations between the biological and social components of health: it reaffirms the centrality of molecular processes for the understanding of phenotypic differences and their mixed biological and social genesis (Moore 2017). Although many strands of contemporary epigenetics dethrone ‘the gene’ from its special causal status as “all-encompassing formative prowess” (Lickliter and Witherington 2017, 127), epigenetic ‘factors’ or ‘markers’ allege to represent and explain this complex organism-environment relationship. Under this reading, epigenetics simply offers a biosocial deterministic ontology in the place of its genetic predecessor—one that reinforces the view taking the (epi)genetic level of analysis as “explanatorily foundational” (Lickliter and Witherington 2017, 127). The difference is that “research on epigenetics […] simply shift[s] attention from one kind of molecule, such as DNA and the “code” it carries”, to another kind of molecular explanation (Moore 2017, 73 cited in Lickliter and Witherington 2017). This is always based on genetic material—chemical modification of DNA or histone proteins and their informational value—but adds to the explanatory value of this molecular substrate the value of analogue to the organism’s relation with experiences and environments. Molecules explaining biography and milieu in epigenetics are thus a distinct biosocial ontology from the thick material-semiotic ontology of the biosocial in humanistic and STS debates. The work towards producing the entanglement of the body and its environments in the language of chemical modifications of DNA and the genome raises therefore several questions about the assumptions, conceptions, entities and constructs that partake to the production of a biosocial ontology of health and disease. The open question is whether the ontologies scientists enact in epigenetics can be reconciled with STS biosocial conceptions of life. In other words, are these distant ontologies of the biosocial incommensurable? In Ludwig Fleck’s jargon of thought styles (Fleck 1981), there is little doubt that epigenetic researchers think and see differently from social scientists the situated character of our biologies. I will elaborate further on the relevance of this question of incommensurability and intercommunication of thought styles in the discussion. Suffice to say here that, in the performative idiom of STS scholarship (e.g. Pickering 2017), the answer to this ontological question heavily depends on

16

L. CHIAPPERINO

the epistemic practices and activities that make up such distant biosocial worlds. Let us turn to this other meaning of ‘biosocial’.

3

Biosocial as Qualifier of Epistemic Practices

The concept of ‘biosocial’ designates also a contested epistemic space at the crossroad of the social and life sciences (Meloni et al. 2018b). Indeed, twentieth-century attempts to bridge social and biological studies of the human condition are fraught with pressing theoretical, methodological and political disagreements. As argued already by Baldwin and Baldwin (1980), sociologists who are willing to consider the relevance of a biological explanation of social phenomena should be wary of the limitations, assumptions and implications that models of knowledge production in biology offer to think about complex looping effects between behaviours, bodies and patterns of social organisation. Taking at issue the paradigmatic case of sociobiology (Wilson 2000; see also Fox Keller 2016), Baldwin and Baldwin point to the irreducible limitations put by molecular biology (its methods, tools and ways of knowledge-making) on the understanding of biosocial processes. Not only, they argue, biologists tend to rely on a limited concept of society—a point reiterated by critics today about the black box of the ‘environment’ in epigenetics (Kuzawa 2017; Pinel 2022)—but the centrality of the gene in sociobiology runs the risk of turning socialisation into a mere “multiplier effect” of genetically controlled traits (Baldwin and Baldwin 1980, 10). Several have looked into whether plasticity thinking in epigenetics offers a less hostile ground for building a shared epistemic space of biosocial research across the life and social sciences (Niewöhner 2015; Müller et al. 2017; Meloni et al. 2018b). However, an integrative biosocial science capturing the complex, non-linear social-biological transitions in the shaping of the epigenome is not much more than a chimaera. The reality of knowledge practices in epigenetics is far more intricate, contested and heterogeneous. On the one hand, it is undeniable that this science pins an open conception of causality against strong views of genetic determinism and reductionism. Different strands of epigenetic research contribute different versions of genotype-to-phenotype (G-to-P) transitions as an interactive phenomenon: being it developmental, environmental or time-sensitive forces, G-to-P in epigenetics is a complex, open-ended process crafting the organism amidst several contributing factors (Pradeu 2016). This feature of epigenetics undeniably opens

THE TURN TOWARDS ‘THE BIOSOCIAL’ IN EPIGENETICS …

17

up biosciences to interdisciplinary collaborations that wish to embed the complexity of social conditions and relations into the life sciences’ measurements of individual health, well-being, risks and predispositions (Niewöhner 2015). On the other hand, the practice of epigenetic science is also grounded on research methods and standards of evidence that leave little room to an integrative biosocial science. Primarily, this can be illustrated by considering the lack of complex measures of social conditions and the environment in epigenetics. In fact, several critics have pointed out that epigenetics only reinterprets social differences as mere molecular changes (Chung et al. 2016). This entails a slippage in interpreting proxy measures of experiences and exposures as biological differences— a process anthropologist Jörg Niewöhner called “the molecularisation of biography and milieu” (2011, 291). A good example may be provided by studies including measurements of the social determinants of health in association with epigenetic analyses. In a systematic review of this literature, Evans and colleagues have found that the operationalisations of the social determinants of health in epigenetics boil down to the inclusion of individual sociodemographic traits (e.g. education, income, ethnicity) in correlational analyses of measurements of epigenetic differences, such as DNA methylation (Evans et al. 2021). The shift towards a biology sensitive to environmental harms, structural social conditions (e.g. disadvantage or discrimination) and developmental trajectories seems therefore far from finding its operationalisation in such experimental work. While this research certainly contributes to an affirmation of the relevance of biosocial mechanisms producing health trajectories, these measures have a “limited ability” to highlight “upstream, systemic factors as key determinants of population health” (Evans et al. 2021, 9). Yet, at a more foundational level, an integrative biosocial science may be incompatible with common epigenetic research practices for a different reason. The point here may be less about the correlates putatively explaining biological-and-social variation in health than about the defining epistemological features of a thick biosocial science of health. Under this interpretation, epigenetic research per se may be unable to fully grasp a thick, holistic view of biosocial processes likely because knowledge of these processes in the social sciences does not fundamentally require the same kind of empirical exercise. Clearly, life scientists today consider epigenetics “a unified framework in which […] multi-modal and multi-scale data can be related, interpreted, and explored” (Coda and Gräff 2020, 5). Epigenetics here retains the strength of quantitative

18

L. CHIAPPERINO

and experimental sciences while leveraging the capacities of informational sciences to promote an integrative approach to health. And yet, one might ask, is this system, data-intensive science of biosocial processes—grounded on what Henrik Vogt and colleagues have called ‘technoscientific holism’ (Vogt et al. 2016)—commensurable to the science of situated biologies and health as unique individual-specific, embodied phenomenon (Niewöhner and Lock 2018)? The mere idea that these complex materialsemiotic processes may be decomposable into distinct biological and social components is a dualist and reductionist framing that is incompatible with the style of reasoning about the biosocial in these strands of social sciences. This cursory look onto the methodological canons of epigenetics hints at several gaps and challenges preventing this field from exploring and embracing the approach to ‘the biosocial’ in STS scholarship. Some of the defining features of those technoscientific practices risk turning distinctively thick human social processes into biological correlates putatively explaining variation in health and disease. And yet, the problem seems not only one of informational value, but also one of incommensurability: the research problems, styles of reasoning, communicative interactions, methods, facts and standard of evidence in this field of contemporary biomedicine fall short of a conception of ‘the biosocial’ grounded on the historical and social contingency of the material body. This is not meant to deny the value of epigenetic representations of biosocial processes: the models, technologies and knowledge practices of the field (e.g. molecular biology, translational medicine, epidemiology) produce a valuable appreciation of the complex dynamics joining material biological processes with experiences, behaviours and social conditions. Yet, recognising the incommensurability of these different styles of biosocial reasoning raises a considerable issue to those regularly engaging in collaborations around epigenetics (Müller et al. 2017; Chiapperino and Paneni 2022). Does this mean that these considerations of incommensurability prevent any interdisciplinary biosocial experimentation? Epigenetics entertains collaborations and a critical view of biology as contingent on material and social environments, such as political, economic and historical factors. It is a field where the research agendas of the life and social sciences—while irreconcilable in fine—can at least try to converge and integrate. This is not meant to say that one can easily shrug off epistemological questions of different styles of reasoning. Plus one should also not underestimate how loaded these collaborations can be: the encounter

THE TURN TOWARDS ‘THE BIOSOCIAL’ IN EPIGENETICS …

19

between different epistemic cultures and traditions of research has its own intricacies, power dynamics, frustrations and risks (Callard and Fitzgerald 2015). Nevertheless, I side with those who, in other contexts (e.g. Olsson et al. 2015), have made a distinction between two distinct types of knowledge integration: pluralism and unification. Issues of incommensurability between the epistemic approaches to the biosocial in life and social sciences only make the ideal of a unified biosocial science unrealistic. However, they do not constitute an obstacle to a pluralistic, problemoriented view of collaboration and interdisciplinary experimentation. If it is true that epigenetics is failing on an objective it has largely given itself (i.e. providing a thicker appreciation of biosocial processes in health) (Chiapperino and Paneni 2022), it is also important to recall that epigenetics does not need to be an interdiscipline to achieve this goal (Dubois et al. 2020). While it is undeniable that the field is wanting in the ways it operationalises the environment, one should not underestimate that a full grasp of the deep socio-environmental embeddedness of our biology is no one’s prerogative—neither of this experimental science, nor of any social science approach for that matter. What, then, are the benefits of expecting a biology of social position to grasp a contingent process of biological and social differentiation that we designate as ‘health’? The point is rather how the environments of epigenetics are “explicitly delineated or contoured for any investigation”, with full awareness that “doing so is a device, with unavoidable limitations” (Lock 2020, 40). If socially patterned exposures and experiences connect with biological changes (and health outcomes) through multiple pathways and iterative loops, this complex biosocial view of health may be put to test in investigative collaborations of populational studies, without necessarily expecting the ultimate biosocial design of research to study them. In synthesis, a frugal, pluralist theoretical stance on practices of interdisciplinary, methodological and interventional experimentation may be required to bring a biosocial agenda around epigenetics beyond issues of incommensurability.

4 Biosocial as an Attribute of Socio-Political Strategies of Intervention Another use of the concept ‘biosocial’ in relation to epigenetics has to do with the ways this knowledge inspires socio-political strategies to govern and intervene into the biological and social determinants of

20

L. CHIAPPERINO

health. Primarily, this understanding of the concept emerged in the 1990s and pointed to genetics as truth-discourse idiosyncratically producing notions of social difference (e.g. ethnicity, disease risks, group identity—see Rabinow 1996; Hacking 2006). In the words of anthropologist Paul Rabinow (1996, 99), “biosociality” designates the emergence of a “new type of autoproduction”—a situated phenomenon of identity- and group-making heavily structured by the possibilities and facts of genetics. Rabinow’s work detailed how genetics crafted new modes of subjectivation and individualisation of health, as well as novel “ways of thinking and acting at the level of population groups and collectivities” (Rabinow and Rose 2006, 204). In the case of epigenetics, these processes are increasingly at play when evidence of the epigenetic effects of social conditions and experiences becomes the explanatory framework for the embodied experience of social conditions. Building upon animal studies of the epigenetic effects of conditioning experiments (Turecki and Meaney 2016), epigenetic biomarkers of psychosocial stress and/or traumatic events are a growing object of research in translational psychiatry (Dahrendorff and Uddin 2021). Notwithstanding the controversial status of this knowledge (Deichmann 2020; Chiapperino and Paneni 2022), these claims about the embodiment of historical discriminations (Warin et al. 2019) or structural disadvantages (Müller and Kenney 2021) increasingly circulate in wider social circles and provide meaning and value to narratives of individual or group identity as well as claims of restorative justice. As Rabinow argued, biosociality is not just identity-making for individuals and/or collectives; the political reach of the concept exceeds these social processes. Biosociality also consists of the production of a specific repertoire of interventions into life, which intertwine with the knowledge practices of the molecular biology lab. Since the advent of genetics and genetic biotechnologies, he argued, nature became “culture understood as practice” (1996, 99); that is, health and bodily factors become amenable to the expert analysis and interventions of the life sciences laboratory. This means, drawing an example from epigenetics, that facts about the epigenetic effects of adversity do not only transform the idea of belonging to an oppressed and/or disadvantaged group. Rather, they produce the alleged biological mechanisms of these processes, they highlight some opportunities to intervene into them (to the detriment of others), and thus, they radically impact the politics and policies around these matters. This crucial element, at least from an STS perspective, reminds us that “representing and intervening, knowledge and power,

THE TURN TOWARDS ‘THE BIOSOCIAL’ IN EPIGENETICS …

21

understanding and reform, are built in, from the start, as simultaneous goals and means” of knowledge-making and technological innovation (Rabinow 1996, 93). In this vein, biosociality in epigenetics can also be defined as follows: the process aligning knowledge-making on biosocial traffic with specific social and political orientations to act upon these phenomena for the benefit of individual and group health. Critical scholarship examining the opportunities for social change and reform afforded by epigenetics has directed its critical eye to the different and rival biosocialities (understood as strategies and repertoires of intervention) weaved into epigenetic knowledge production (Niewöhner 2011; Chiapperino and Testa 2016; Kenney and Müller 2017; Chiapperino 2021). While epigenetics’ potential for emancipation and reparation of injustices is undeniable, these same analyses have often questioned the actual contribution of epigenetics to human flourishing, health equity and citizen empowerment (Chiapperino and Testa 2016). The way political views and technoscientific strategies of intervention (e.g. pharmacological, environmental, behavioural, clinical, populational) intersect in epigenetic research is in fact questionable. While discourses on the necessity for complex social interventions preventing epigenetic risks abound in the field (Penkler 2022), the repertoire of intervention built into epigenetic science is far more simplistic (Chiapperino 2021; Chiapperino and Paneni 2022). The most frequent approaches to interventions into the epigenetic marks of social and environmental exposures consist of behavioural interventions or specific classes of drugs to act on the reversibility of the epigenetic machinery. At times, active molecules such as DNA methyltransferases (DNMTs) and histone deacetylases (HDACs) inhibitors can be combined with behavioural therapy to intervene into (reverse) the traces of (adverse) experiences and exposures (Szyf 2009; Day 2014). These approaches advance a biomedical and/or individualistic mode of intervention grounded on pharmacology, which propagates a distinctive biosociality of epigenetics. On the one hand, pharmacological interventions translate everyday social contexts and significant biographical events into matters to be dealt with at the molecular level (Niewöhner 2011). On the other hand, behavioural interventions treat a process with far-reaching psychological, social and environmental ramifications as an individual matter—and one that, notwithstanding such ramifications, does not extend beyond the individual body (Landecker 2016).

22

L. CHIAPPERINO

But what does go missing in these operationalisations of intervention into disease and the epigenome? Some practitioners of the field, as well as most social/humanistic critics (see Chiapperino 2018 for a review), agree that epigenetics should activate a different kind of biosociality. The field’s focus on the multifactorial aetiology of disease conditions calls for a different mode of intervention: one that considers the biological, psychological social components of epigenetic differences thoroughly, meaning that it does not obliterate the relevance of psychotherapy much like any relational, socio-economic or structural intervention into the epigenome and health (Kular and Kular 2019). In fact, little is done today in epigenetics to fully capitalise on the syndemic dimensions of these biological differences and disease conditions (Singer et al. 2017); that is, to activate the health promotion potential of upstream social, political, and structural determinants of health differences readable in the epigenome. Drawing from notable exceptions to this general trend—which can be found in the work of critical public health researchers such as Arline T. Geronimus (2013), Gene H. Brody (Brody et al. 2016) or neuroscientists like Bruce McEwen (2017)—this section will end on a suggestion about the potential biosocialities of epigenetics that are yet to be built into its repertoire of interventions. As I have documented elsewhere (2021; see also Chiapperino and Paneni 2022), actors in the field are aware of these criticisms. Modelling complex experiences/exposures and treating them as druggable biological differences in the epigenome is often an approach that is convenient, but far from being fully satisfying for practitioners too. In a laboratory study (Chiapperino 2021), I detailed such discontent and illustrated how the need to cultivate different biosocialities of epigenetics goes beyond rhetoric: also experimental scientists explore environmental interventions into the ‘aberrant’ epigenetic marks of early-life stress. Specifically, my argument built on a proof-of-principle paper—the first of this kind in behavioural epigenetics—which used an enriched environment promoting physical, cognitive and interactional stimulation for mice conditioned with chronic early-life stressors. This environmental intervention was able to reverse the epigenetic marks of early-life stress in a mouse model, as well as their behavioural correlates, thus preventing transmission of these modifications across generations. In this regard, these experimental practices enacted an interventional strategy alternative to pharmacological treatments of stress-related conditions in epigenetics (Szyf 2009)—in other words, they illustrated that a different biosociality of epigenetics is not

THE TURN TOWARDS ‘THE BIOSOCIAL’ IN EPIGENETICS …

23

only a prerogative of critique and/or discourse, but rather it can be weaved into the lab practices of the field too. Moving beyond individualoriented and/or pharmacological interventions, the environment (even if just as modelled and standardised living conditions of experimental animals) was here activated to tinker with the reversibility and malleability of the epigenome. At one level, this example of an environmental intervention into the epigenome illustrates how scientists render complex social-biological and/or body-environment transitions amenable to the practical affordances of the laboratory (Chiapperino 2021). As such, it is an interventional strategy fully inscribed in the ontological and epistemological style of reasoning about the biosocial which I criticised in the two previous sections. Yet, while still liable to critique, this interventional strategy also feeds into a biosociality that converges with social sciences and humanities orientations. In a nutshell, these experimental practices promote a normative translation of epigenetics, which differs substantially from the one of the pharmacological and biomedical mode of intervention, and that may be commensurable to the one promoted by social sciences and humanities critics. For instance, while they certainly provide a molecular ontology of biosocial processes, environmental interventions (also when restricted to animal models) promote the public and political salience of context and living conditions in these processes. The brain and the epigenome’s plasticity are here, perhaps, still problematic ontologies of complex biosocial, embodied and situated processes such as stress-related diseases (i.e. they still reduce to molecular differences health conditions with major social and environmental ramifications). Yet, the normative relevance of these experimental variables is that of molecular endpoints of social and/or environmental change instead of simply druggable molecular mechanisms disjoined from broader “social intervention” (cf. McEwen 2022, 2). Furthermore, this kind of experiments support the view that structural change in our society, more than just individual treatments, should be part of the tested repertoires of intervention of epigenetics. While, again, these practices do not lack limitations (cf. Chiapperino 2021), they break at least a problematic equation between what is “believed to be feasible” as intervention and what the institutionalisations of epigenetics should entail, between the evidence of intervention available and the fundamental “political issue[s]” epigenetics illuminates (Geronimus 2013, S61). If it is true that most of the molecular evidence on the

24

L. CHIAPPERINO

reversal of epigenetic marks is produced with the use of drugs, environmental intervention studies remind us that it is also pertinent to consider the efficacy of social and environmental interventions—even if these are less appealing, practical and practised in experimentation. Finally, and in a more straightforward way, environmental interventions are just a reminder that non-biological modulators of health can be factors of resilience and/or therapeutic intervention into biological vulnerabilities. These studies activate the continuity between the social and the biological, the body and the environment, the molecular and the contextual also in the interventional repertoire of epigenetics. Far from being only a concern of public health scientists (Brody et al. 2016), the epigenome’s responsiveness to contextual interventions should not be “deep[ly] unfamiliar” to the popular methods of this science (Geronimus 2013, S57). Under this reading, experiments exploring environmental interventions into the epigenome show that a biosociality made of structural interventions and social change (instead of individual and molecular interventions) is not just the prerogative of social sciences’ normative reasoning about ‘the biosocial’. While the ontology and episteme attached to these practices are still the one I criticised in the previous sections, their ensuing normative translations may constitute a commensurable ground of collaboration knowledge practices about the biosocial with a thoroughly different epistemic and ontological orientation.

5

Conclusions

Fleshing out the ontological, epistemic and socio-political dimensions of ‘the biosocial’ reveals its material-semiotic alternative realities across the life and social sciences. I started with characterising these different dimensions of the concept in STS and humanistic circles; then, I probed how these ideas and ways of knowing biosocial processes are at play in the experimental work of epigenetics. Life scientists’ approach to the biosocial entanglements at the basis of health and disease is characterised by a dynamic inclusion and exclusion of different bits of ‘the biosocial’, as complex situated ontology, knowledge practice, or even social and political view of health. These distinct approaches to ‘the biosocial’ are, in other words, “assemblages of relations” that do the “realities” of these processes differently for science and society (Law 2009, 2). The recognition of such alternative realities of ‘the biosocial’ calls for a reflection on their in-commensurability. Biosocial ontologies, epistemic

THE TURN TOWARDS ‘THE BIOSOCIAL’ IN EPIGENETICS …

25

practices studying social-biological hybridities in health or even sociopolitical interventions into this nexus are fundamentally contingent. They are assembled in specific local and momentary circumstances, within traditions and configurations of research that differ substantially. In a nutshell, they are the result of specific styles of reasoning about the biosocial (Hacking 1992; Sciortino 2017). While generating representations, facts and normative injunctions around the importance of social-biological transitions in health differentiation, these styles of biosocial reasoning remind us that it may be difficult to bring together different communities of scholars around a unified standard of evidence, language and/or approach to biosocial processes of health differentiation. There is not, in all likelihood, an independent, singular or coherent way to dissect biosocial processes that is in easy reach to collaborations around epigenetics and social-biological transitions in health. Whether one looks at the processual ontology of biosocial becoming in the social sciences (cf. Ingold and Pálsson 2013), or to the understanding of biosocial traffic as molecular correlates of experiences in epigenetics (cf. Lickliter and Witherington 2017), it is multiplicity which stands out when analysing the alternative realities of ‘the biosocial’ across the life and social sciences. The first conclusion of the present chapter is therefore a call to seriously consider the richness of biosocial ideas, their incommensurability and, perhaps, abandon the chimera of a unified style of biosocial reasoning. It may be more productive to employ this epistemological question as analytical tool to dissect the multiple standards of objectivity and process ontologies of health at stake in the practices that produce and perform biosocial processes in different disciplinary settings. This leads me to a second consideration building on the performativity and incommensurability of alternative styles of biosocial reasoning. Far from being a stall, this recognition should be read as an invitation to put practice before theory, to prioritise experimentation and social change in the face of theory-making and epistemological unification. A dynamic, non-gene-centric, environmentally embedded and reactive view of health may be incommensurable to the methodological reductionism of the life sciences. Conversely, moving into a molecularly informed biosocial science may require analytical, statistical and quantitative sensitivities currently alien to the worldviews of most STS scholars (Meloni et al. 2018a). Does that mean that no communication among these styles and though collectives of the biosocial is possible (Fleck 1981)? As shown

26

L. CHIAPPERINO

in the previous sections, there might be circumstances where the differences between ontologies and epistemic approaches to the biosocial are great. Yet, we have seen, such distance may be less important when it comes to few of the biosocialities these different styles produce. At least in part—however small it may be—life and social scientists converge on specific social and political circulations of biosocial facts. There exists a style of reasoning about the socio-political and public health role of epigenetic knowledge—likely mediated by traditions of research in social epidemiology (see Louvel and Soulier 2022)—that converges with longstanding social sciences commitment to tackle the social gradient of health before pathologising and molecularising social conditions. It is in such liminal overlaps and comparisons that, by paraphrasing Ludwig Fleck, “the divergence” between “thought styles” of biosocial reasoning could “dwindle into nothing” (Fleck 1981, 108). While still diverging as ontologies or knowledge repertoires of the biosocial, practices insisting on the importance of social interventions share a potential normative translation and channel of communication about biosocial facts: the one insisting on social, structural change, life conditions and contexts of living as actionable means of intervention into the biosocial processes of health differentiation. When it comes to its socio-political dimension, the life scientists and social scientists sharing this recognition might be practising only “nuances of style, varieties in style” (Fleck 1981, 108). In other words, a task may be in order for the critic who goes beyond ontological and epistemological questions of incommensurability: this may be to let converging forms of socio-political change be visible across distant styles of biosocial experimentation. In practice, this would mean mapping the multiple relations between tools, facts, methods and politics which get built into specific epistemic practices of biosocial research. What concretely different resources for an alternative social circulation of biosocial facts exist within the practical repertoire of epigenetics? With what inherent resemblance to the political undertones of the biosocial in STS? Beyond the chimaera of an integrative science lies the opportunity for acting counter the unequal socio-material ecologies of life which are a shared matter of concern to distant thought collectives.

THE TURN TOWARDS ‘THE BIOSOCIAL’ IN EPIGENETICS …

27

Bibliography Baldwin, John D., and Janice I. Baldwin. 1980. ‘Sociobiology or Balanced Biosocial Theory?’ The Pacific Sociological Review 23 (1): 3–27. https://doi.org/ 10.2307/1388800. Barad, Karen. 1998. ‘Getting Real: Technoscientific Practices and the Materialization of Reality’. Differences: A Journal of Feminist Cultural Studies 10 (2): 87–91. Bird, Adrian. 2007. ‘Perceptions of Epigenetics’. Nature 447 (7143): 396–98. https://doi.org/10.1038/nature05913. Brody, Gene H., Tianyi Yu, and Steven R. H. Beach. 2016. ‘Resilience to Adversity and the Early Origins of Disease’. Development and Psychopathology 28 (4 Pt 2): 1347–65. https://doi.org/10.1017/S0954579416000894. Callard, Felicity, and Des Fitzgerald. 2015. Rethinking Interdisciplinarity Across the Social Sciences and Neurosciences. Basingstoke, Hampshire and New York, NY: Palgrave Macmillan. https://www.palgrave.com/gp/book/978113740 7955. Cavalli, Giacomo, and Edith Heard. 2019. ‘Advances in Epigenetics Link Genetics to the Environment and Disease’. Nature 571 (7766): 489–99. https://doi.org/10.1038/s41586-019-1411-0. Cerutti, Janine, Alexandre A. Lussier, Yiwen Zhu, Jiaxuan Liu, and Erin C. Dunn. 2021. ‘Associations between Indicators of Socioeconomic Position and DNA Methylation: A Scoping Review’. Clinical Epigenetics 13 (1): 221. https://doi.org/10.1186/s13148-021-01189-0. Chiapperino, Luca. 2018. ‘Epigenetics: Ethics, Politics, Biosociality’. British Medical Bulletin 128 (1): 49–60. https://doi.org/10.1093/bmb/ldy033. ———. 2021. ‘Environmental Enrichment: An Experiment in Biosocial Intervention’. BioSocieties 16 (1): 41–69. https://doi.org/10.1057/s41292-01900181-5. Chiapperino, Luca, and Francesco Paneni. 2022. ‘Why Epigenetics Is (Not) a Biosocial Science and Why That Matters’. Clinical Epigenetics 14 (144): 6. https://doi.org/10.1186/s13148-022-01366-9. Chiapperino, Luca, and Francesco Panese. 2019. ‘On the Traces of the Biosocial: Historicizing “Plasticity” in Contemporary Epigenetics’. History of Science, 59 (1), 3–44. https://doi.org/10.1177/0073275319876839. Chiapperino, Luca, and Giuseppe Testa. 2016. ‘The Epigenomic Self in Personalised Medicine: Between Responsibility and Empowerment’. The Sociological Review, 64 (1_suppl), 203–220. https://doi.org/10.1111/2059-7932. 12021. Chung, Emma, John Cromby, Dimitris Papadopoulos, and Cristina Tufarelli. 2016. ‘Social Epigenetics: A Science of Social Science?’ The Sociological Review Monographs 64 (1): 168–85. https://doi.org/10.1002/2059-7932.12019.

28

L. CHIAPPERINO

Coda, Davide Martino, and Johannes Gräff. 2020. ‘Neurogenetic and Neuroepigenetic Mechanisms in Cognitive Health and Disease’. Frontiers in Molecular Neuroscience, 13. https://doi.org/10.3389/fnmol.2020.589109. Dahrendorff, Jan, and Monica Uddin. 2021. ‘Epigenetic Epidemiology of Psychiatric Disorders’. In Epigenetics in Psychiatry, edited by Jacob Peedicayil, Dennis R. Grayson and Dimitrios Avramopoulos, Academic Press, Elsevier, 11–142. Darling, Katherine Weatherford, Sara L. Ackerman, Robert H. Hiatt, Sandra Soo-Jin Lee, and Janet K. Shim. 2016. ‘Enacting the Molecular Imperative: How Gene-Environment Interaction Research Links Bodies and Environments in the Post-Genomic Age’. Social Science & Medicine 155 (April): 51–60. https://doi.org/10.1016/j.socscimed.2016.03.007. Day, Jeremy J. 2014. ‘New Approaches to Manipulating the Epigenome’. Dialogues in Clinical Neuroscience 16 (3): 345–57. Deans, Carrie, and Keith A. Maggert. 2015. ‘What Do You Mean, “Epigenetic”?’ Genetics 199 (4): 887–96. https://doi.org/10.1534/genetics.114.173492. Deichmann, Ute. 2020. ‘The Social Construction of the Social Epigenome and the Larger Biological Context’. Epigenetics & Chromatin 13 (1): 37. https:// doi.org/10.1186/s13072-020-00360-w. Dubois, Michel, Séverine Louvel, and Emmanuelle Rial-Sebbag. 2020. ‘Epigenetics as an Interdiscipline? Promises and Fallacies of a Biosocial Research Agenda’. Social Science Information 59 (1): 3–11. https://doi.org/10.1177/ 0539018420908233. Evans, Linnea, Michal Engelman, Alex Mikulas, and Kristen Malecki. 2021. ‘How Are Social Determinants of Health Integrated into Epigenetic Research? A Systematic Review’. Social Science & Medicine 273 (March): 113738. https://doi.org/10.1016/j.socscimed.2021.113738. Fleck, Ludwik. 1981. Genesis and Development of a Scientific Fact, edited by Thaddeus J. Trenn, Robert K. Merton, and Thomas S. Kuhn. Translated by Frederick Bradley. New edition. Chicago (USA): University of Chicago Press. Fox Keller, Evelyn. 2010. The Mirage of a Space Between Nature and Nurture. Durham, NC: Duke University Press. ———. 2015. ‘The Postgenomic Genome’. In Postgenomics: Perspectives on Biology After the Genome, edited by Sarah S. Richardson and Hallam Stevens. Duke University Press. https://doi.org/10.1215/9780822375449-002. ———. 2016. ‘Thinking About Biology and Culture: Can the Natural and Human Sciences Be Integrated?’ The Sociological Review, 64 (1_suppl): 26–41. https://doi.org/10.1111/2059-7932.12011. Geronimus, Arline T. 2013. ‘Deep Integration: Letting the Epigenome Out of the Bottle Without Losing Sight of the Structural Origins of Population Health’. American Journal of Public Health 103 (1): 56–63. https://doi. org/10.2105/AJPH.2013.301380.

THE TURN TOWARDS ‘THE BIOSOCIAL’ IN EPIGENETICS …

29

Hacking, Ian. 1992. ‘“Style” for Historians and Philosophers’. Studies in History and Philosophy of Science Part A 23 (1): 1–20. https://doi.org/10.1016/ 0039-3681(92)90024-Z. ———. 2006. ‘Genetics, Biosocial Groups & The Future of Identity’. Daedalus 135 (4): 81–95. Haraway, Donna. 1990. Simians, Cyborgs, and Women: The Reinvention of Nature. New York: Routledge. Horsthemke, Bernhard. 2022. ‘A Critical Appraisal of Clinical Epigenetics’. Clinical Epigenetics 14 (1): 95. https://doi.org/10.1186/s13148-022-013 15-6. Ingold, Tim, and Gísli Pálsson, eds. 2013. Biosocial Becomings: Integrating Social and Biological Anthropology. New York: Cambridge University Press. Jablonka, Eva, and Marion J. Lamb. 2005. Evolution in Four Dimensions: Genetic, Epigenetic, Behavioral, and Symbolic Variation in the History of Life. Life and Mind. Cambridge, Mass: MIT Press. Jaenisch, R., and A. Bird. 2003. ‘Epigenetic Regulation of Gene Expression: How the Genome Integrates Intrinsic and Environmental Signals’. Nature Genetics 33 (March): 245–54. https://doi.org/10.1038/ng1089. Kenney, Martha, and Ruth Müller. 2017. ‘Of Rats and Women: Narratives of Motherhood in Environmental Epigenetics’. BioSocieties 12 (1): 23–46. https://doi.org/10.1057/s41292-016-0002-7. Kular, Sonia, and Lara Kular. 2019. ‘Epigenetics as a Partner of Psychiatry: Toward a Personalized Approach of the Patient’. Evolution Psychiatrique 84 (1): 207–21. https://doi.org/10.1016/j.evopsy.2018.05.003. Kuzawa, Christopher W. 2017. ‘Which Environments Matter in Studies of Early Life Developmental Plasticity?’ Evolution, Medicine, and Public Health 2017 (1): 188–90. https://doi.org/10.1093/emph/eox024. Landecker, Hannah. 2016. ‘The Social as Signal in the Body of Chromatin’. The Sociological Review Monographs 64 (1): 79–99. https://doi.org/10.1002/ 2059-7932.12014. Landecker, Hannah, and Aaron Panofsky. 2013. ‘From Social Structure to Gene Regulation, and Back: A Critical Introduction to Environmental Epigenetics for Sociology’. Annual Review of Sociology 39 (1): 333–57. https://doi.org/ 10.1146/annurev-soc-071312-145707. Law, John. 2011. ‘Collateral Realities’. In The Politics of Knowledge, edited by Patrick Baert, Fernando Dominguez Rubio, London: Routledge. Lickliter, Robert, and David C. Witherington. 2017. ‘Towards a Truly Developmental Epigenetics’. Human Development 60 (2–3): 124–38. https://doi. org/10.1159/000477996. Lock, Margaret. 2013. ‘The Epigenome and Nature/Nurture Reunification: A Challenge for Anthropology’. Medical Anthropology 32 (4): 291–308. https://doi.org/10.1080/01459740.2012.746973.

30

L. CHIAPPERINO

———. 2015. ‘Comprehending the Body in the Era of the Epigenome’. Current Anthropology 56 (2): 151–77. https://doi.org/10.1086/680350. ———. 2020. ‘Permeable Bodies and Environmental Delineation’. In Biosocial Worlds, edited by Jens Seeberg, Andreas Roepstorff, and Lotte Meinert, University College London Press. 15–43. Louvel, Séverine, and Alexandra Soulier. 2022. ‘Biological Embedding vs. Embodiment of Social Experiences: How These Two Concepts Form Distinct Thought Styles Around the Social Production of Health Inequalities’. Social Science & Medicine 314 (December): 115470. https://doi.org/10.1016/j. socscimed.2022.115470. McEwen, Bruce S. 2017. ‘Allostasis and the Epigenetics of Brain and Body Health Over the Life Course: The Brain on Stress’. JAMA Psychiatry 74 (6): 551–52. https://doi.org/10.1001/jamapsychiatry.2017.0270. McEwen, Craig A. 2022. ‘Connecting the Biology of Stress, Allostatic Load and Epigenetics to Social Structures and Processes’. Neurobiology of Stress 17 (March): 100426. https://doi.org/10.1016/j.ynstr.2022.100426. Meloni, Maurizio. 2016. Political Biology. Basingstoke, Hampshire and New York, NY: Palgrave Macmillan. Meloni, Maurizio, John Cromby, Des Fitzgerald, and Stephanie Lloyd. 2018a. ‘Introducing the New Biosocial Landscape’. In The Palgrave Handbook of Biology and Society, London: Palgrave Macmillan. 1–22. ———, eds. 2018b. The Palgrave Handbook of Biology and Society. London: Palgrave Macmillan. Meloni, Maurizio, Simon J. Williams, and Paul Martin, eds. 2016. Biosocial Matters: Rethinking the Sociology-Biology Relations in the Twenty-First Century. Sociological Review Monograph. Chichester, West Sussex and Malden, MA: Wiley-Blackwell. http://eu.wiley.com/WileyCDA/WileyTitle/ productCd-1119236517.html. Montagu, Ashley. 1956. The Biosocial Nature of Man. Grove Press. Moore, David S. 2017. ‘The Potential of Epigenetics Research to Transform Conceptions of Phenotype Development’. Human Development 60 (2–3): 69–80. https://doi.org/10.1159/000477992. Morange, Michel. 2018. ‘The Historiography of Molecular Biology’. In Handbook of the Historiography of Biology, edited by Michael Dietrich, Mark Borrello, and Oren Harman, Cham: Springer International Publishing, 1-20. https://doi.org/10.1007/978-3-319-74456-8_11-1. Müller, Ruth, Clare Hanson, Mark Hanson, Michael Penkler, Georgia Samaras, Luca Chiapperino, John Dupré, et al. 2017. ‘The Biosocial Genome?: Interdisciplinary Perspectives on Environmental Epigenetics, Health and Society’. EMBO Reports, September, e201744953. https://doi.org/10.15252/embr. 201744953.

THE TURN TOWARDS ‘THE BIOSOCIAL’ IN EPIGENETICS …

31

Müller, Ruth, and Martha Kenney. 2021. ‘A Science of Hope? Tracing Emergent Entanglements between the Biology of Early Life Adversity, Trauma-Informed Care, and Restorative Justice’. Science, Technology, & Human Values 46 (6): 1230–60. https://doi.org/10.1177/0162243920974095. Nanney, D. L. 1958. ‘Epigenetic Control Systems*’. Proceedings of the National Academy of Sciences of the United States of America 44 (7): 712–17. Nicholson, Daniel J., and John Dupré, eds. 2018. Everything Flows: Towards a Processual Philosophy of Biology. Oxford, UK: Oxford University Press. Niewöhner, Jörg. 2011. ‘Epigenetics: Embedded Bodies and the Molecularisation of Biography and Milieu’. BioSocieties 6 (3): 279–98. https://doi.org/ 10.1057/biosoc.2011.4. ———. 2015. ‘Epigenetics: Localizing Biology through Co-Laboration’. New Genetics and Society 34 (2): 219–42. https://doi.org/10.1080/14636778. 2015.1036154. ———. 2020. ‘Situating Biologies: Studying Human Differentiation as MaterialSemiotic Practice’. In Biosocial Worlds: Anthropology of Health Environments beyond Determinism, edited by Jens Seeberg, Andreas Roepstorff, and Lotte Meinert. UCL Press, 44–68. Niewöhner, Jörg, and Margaret Lock. 2018. ‘Situating Local Biologies: Anthropological Perspectives on Environment/Human Entanglements’. BioSocieties, 13, 681–697. https://doi.org/10.1057/s41292-017-0089-5. Olsson, Lennart, Anne Jerneck, Henrik Thoren, Johannes Persson, and David O’Byrne. 2015. ‘Why Resilience Is Unappealing to Social Science: Theoretical and Empirical Investigations of the Scientific Use of Resilience’. Science Advances 1 (4): e1400217. https://doi.org/10.1126/sciadv.1400217. Oyama, Susan. 2000. ‘Causal Democracy and Causal Contributions in Developmental Systems Theory’. Philosophy of Science 67 (S3): S332–47. https://doi. org/10.1086/392830. Penkler, Michael. 2022. ‘Caring for Biosocial Complexity. Articulations of the Environment in Research on the Developmental Origins of Health and Disease’. Studies in History and Philosophy of Science 93: 1–10. https://doi. org/10.1016/j.shpsa.2022.02.004. Peterson, Erik L. 2017. The Life Organic: The Theoretical Biology Club and the Roots of Epigenetics. Pittsburgh, PA: University of Pittsburgh Press. Pickering, Andrew. 2017. ‘The Ontological Turn: Taking Different Worlds Seriously’. Social Analysis 61 (2): 134–50. https://doi.org/10.3167/sa.2017. 610209. Pigliucci, Massimo. 2001. Phenotypic Plasticity: Beyond Nature and Nurture. Baltimore: Johns Hopkins University Press. Pinel, Clémence. 2022. ‘What Counts as the Environment in Epigenetics? Knowledge and Ignorance in the Entrepreneurial University’. Science as

32

L. CHIAPPERINO

Culture 31 (3): 311–33. https://doi.org/10.1080/09505431.2022.204 3840. Pitts-Taylor, Victoria. 2019. ‘Neurobiologically Poor? Brain Phenotypes, Inequality, and Biosocial Determinism’. Science, Technology, & Human Values 44 (4): 660–85. https://doi.org/10.1177/0162243919841695. Pradeu, T. 2016. ‘Toolbox Murders: Putting Genes in Their Epigenetic and Ecological Contexts’. Biology & Philosophy 31 (1): 125–42. https://doi.org/ 10.1007/s10539-014-9471-x. Rabinow, Paul. 1996. Essays on the Anthropology of Reason. Princeton Studies in Culture/Power/History. Princeton, NJ: Princeton University Press. Rabinow, Paul, and Nikolas Rose. 2006. ‘Biopower Today’. BioSocieties 1 (2): 195–217. https://doi.org/10.1017/S1745855206040014. Rheinberger, Hans-Jörg, and Staffan Müller-Wille. 2018. The Gene: From Genetics to Postgenomics. Translated by Adam Bostanci. Expanded, Revised edizione. Chicago: University of Chicago Press. Richardson, Sarah S., and Hallam Stevens, eds. 2015. Postgenomics: Perspectives on Biology after the Genome. Durham: Duke University Press. Rose, Nikolas. 2013. ‘The Human Sciences in a Biological Age’. Theory, Culture & Society 30 (1): 3–34. https://doi.org/10.1177/026327641245 6569. Sady, Wojciech. 2021. ‘Ludwik Fleck’. In The Stanford Encyclopedia of Philosophy, edited by Edward N. Zalta, Winter 2021. Metaphysics Research Lab, Stanford University. https://plato.stanford.edu/archives/win2021/entries/fleck/. Sciortino, Luca. 2017. ‘On Ian Hacking’s Notion of Style of Reasoning’. Erkenntnis 82 (2): 243–64. https://doi.org/10.1007/s10670-016-9815-9. Singer, Merrill, Nicola Bulled, Bayla Ostrach, and Emily Mendenhall. 2017. ‘Syndemics and the Biosocial Conception of Health’. The Lancet 389 (10072): 941–50. https://doi.org/10.1016/S0140-6736(17)30003-X. Stotz, Karola, and Paul Griffiths. 2016. ‘Epigenetics: Ambiguities and Implications’. History and Philosophy of the Life Sciences 38 (4): 22. https://doi.org/ 10.1007/s40656-016-0121-2. Szyf, Moshe. 2009. ‘Epigenetics, DNA Methylation, and Chromatin Modifying Drugs’. Annual Review of Pharmacology and Toxicology 49 (1): 243–63. https://doi.org/10.1146/annurev-pharmtox-061008-103102. Turecki, Gustavo, and Michael J. Meaney. 2016. ‘Effects of the Social Environment and Stress on Glucocorticoid Receptor Gene Methylation: A Systematic Review’. Biological Psychiatry, 79 (2): 87–96. https://doi.org/10.1016/j.bio psych.2014.11.022 Vogt, Henrik, Bjørn Hofmann, and Linn Getz. 2016. ‘The New Holism: P4 Systems Medicine and the Medicalization of Health and Life Itself’. Medicine, Health Care, and Philosophy 19: 307–23. https://doi.org/10.1007/s11019016-9683-8.

THE TURN TOWARDS ‘THE BIOSOCIAL’ IN EPIGENETICS …

33

Waddington, Conrad H. 2012. ‘The Epigenotype’. International Journal of Epidemiology 41 (1): 10–13. https://doi.org/10.1093/ije/dyr184. Warin, Megan, Emma Kowal, and Maurizio Meloni. 2019. ‘Indigenous Knowledge in a Postgenomic Landscape: The Politics of Epigenetic Hope and Reparation in Australia’. Science, Technology, & Human Values, 45 (1) 87–111. https://doi.org/10.1177/0162243919831077. Wilson, Edward O. 2000. Sociobiology: The New Synthesis, Twenty-Fifth Anniversary Edition. Cambridge, MA: Harvard University Press.

Socio-Markers and Information Transmission Federica Russo

1

Bio-Markers and the Molecular Turn in the Health Sciences

Epidemiology, in its various sub-fields, has been and still is a pillar in the generation of evidence in the biomedical sciences and public health. We owe to the hard work of epidemiologists to the last century if we have a good grip on the aetiology of numerous health conditions, from cancers to obesity, from cardiovascular diseases to reproductive health. Results obtained in the field have been groundbreaking and yet often object of controversies. Since the 1970s, epidemiology has seen an important shift at the conceptual and methodological level (Eybpoosh et al. 2017; Honardoost et al. 2018; Schulte and Perera 1993; Vineis and Chadeau-Hyam 2011; Vineis and Perera 2007). Molecular epidemiology

F. Russo (B) Department of Philosophy & ILLC, University of Amsterdam, Amsterdam, The Netherlands e-mail: [email protected] Department of Science and Technology Studies, University College London, London, England

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 É. Giroux et al. (eds.), Integrative Approaches in Environmental Health and Exposome Research, https://doi.org/10.1007/978-3-031-28432-8_3

35

36

F. RUSSO

studies key characteristics of (aggregations of) individuals and of the environment, reaching an unprecedented granularity in the measurements, thanks to the analysis of biological specimens at the molecular level (Rothman et al. 2011). Molecular epidemiology thus marks a new period of epidemiology, in which bio-markers play a central role in the study. In the scientific and philosophical literature, there isn’t just one definition of bio-marker that scholars unanimously accept or endorse, but usually definitions share some common points (Califf 2018; International Programme on Chemical Safety 2001; NCI Dictionaries 2022; Strimbu and Tavel 2010; Yousef et al. 2009; Biomarkers Definitions Working Group 2001): • Bio-markers point to bio-chemical characteristics; • These characteristics are measurable somehow (e.g. with analysis of biospecimens at various omic levels); • These measurements should give insights into normal/pathogenic process; • The purposes of measurement may be different (e.g. diagnosis, monitoring, prediction). The promise of bio-markers is therefore to establish a link between exposure and clinical outcomes, to gain understanding of normal/pathological processes, and in this way to improve on early prediction of clinical outcomes. Bio-markers are used to ‘mark’ bio-chemical processes at different levels (e.g. genomics, proteomics, metabolomics, etc.) and thus to understand how exposure works at the level of molecules, trying to get a grip on the process that leads to disease onset and to clinical conditions. It is important to note from the start that while there is a lot of causal reasoning at work in bio-markers research, it is wrong to equate biomarkers with causes. Bio-markers, as the word suggests, are markers of various kind of processes, and these processes, linking exposure to disease, are causal in character. In these processes, bio-markers may coincide with causes, but not necessarily. This significant conceptual aspect was pointed out as early as 1993 by Paul Schulte: Does the marker represent an event, is it an event itself, is it a correlate of the event, or is it a predictor of the event? The answers to these questions may affect who is sampled, how and when they are sampled, and what confounders or effect modifiers are considered. […]

37

SOCIO-MARKERS AND INFORMATION TRANSMISSION

Thus, a biologic marker often refers to the use made of a piece of biologic information rather than to a specific type of information. (Schulte 1993, 14–15)

Thus, bio-markers cannot be automatically equated with causes. A biomarker may mark a causal factor directly or can mark in some way a biochemical process in which the causal factor is active. I summarize some of the relations between bio-markers and causes with the diagram in Fig. 1 (adapted from Ghiara and Russo [2019]). In 1(a), bio-marker B is part of the causal process linking the cause X and the effect Y. In (b), the bio-marker B and the effect Y are correlated because both are caused by X. In (c), the bio-marker B is correlated with an event Z, that is part of the process from X to Y. In (d), B is just correlated with X and is not causally linked to the effect Y, in which case B may be, for instance, a background condition. Epidemiological studies employing bio-markers do not obtain full knowledge of (the aetiology of) a given health condition or disease; with bio-markers, we aim to understand the key and relevant characteristics of the process of disease, from exposure to clinical conditions. This is in line with recent approaches in cancer research, that focus on the identification of key characteristics of the carcinogenic process, and bio-markers are often used to pinpoint salient moments of mutations (Smith et al. 2016). B Y

B

X

Y

X

1(b)

1(a) B X

Z

Y

X

B

1(c) Fig. 1 Relations between bio-markers and causes

1(d)

Y

38

F. RUSSO

No matter how much the molecular turn has improved our understanding of processes of health and disease, there is a real danger of over-biologizing them. With bio-markers, we manage to understand quite a lot about the bio-chemistry of health and disease. But we need to climb up the ladder again, so to speak, and to relate the biological and the social spheres of health and disease (Blane et al. 2013; Kelly et al. 2014; Kelly and Russo 2017; Kelly-Irving and Delpierre 2017; Kelly-Irving et al. 2015; Vineis et al. 2020; Vineis and Kelly-Irving 2019). The connection between ‘the biological’ and ‘the social’ has been put at the centre of recent projects in the ‘exposome’ field, and in particular of the Lifepath project (d’Errico et al. 2017; Vineis, Avendano-Pabon, et al. 2017; Vineis and Kelly-Irving 2019; Berger et al. 2019; Vineis et al. 2020). However, this is easier said than done, because it is genuinely difficult to reconcile and integrate very different vocabularies, concepts, and methods for studying the phenomena of health and disease across the biomedical and social sciences. It is important to note that some bio-markers are currently used to grasp the effects of some social condition at the bio-chemical level, for instance in studies of chronic stress. While this is clearly an important line of research, this is not what is intended in this paper with the term ‘socio-marker’. A socio-marker, as I shall also explain in detail in Sect. 2, intends to mark a social process, not the effects of a social process on a bio-chemical process, which is a complementary exercise to what is here proposed. In this paper, I make a modest attempt to contribute to the bridging the two realms in the study of health and disease—the social and the biological—by developing the concept of socio-marker, which is meant to complement that of bio-marker. A good starting point in grasping what socio-markers may be is through the more familiar concept of ‘social determinant’. Socio-economic status and education are widely recognized important factors impacting health and disease of individuals. However, as I explain in Sect. 2, we need to use a concept that is finer-grained than that of social determinant. The overall aim of the paper is to offer a general conceptualization of ‘marker’ that can encompass both socio- and bio-markers and that justifies the need to introducing the novel concept of ‘socio-marker’, besides ‘social determinant’. In Sect. 2, I introduce the concept of socio-marker to complement that of bio-marker. I explain the way in which introducing socio-makers can help us study the ‘social’ part of health and disease at a level of granularity that is comparable to bio-markers study. I also compare the concept of socio-marker with that

SOCIO-MARKERS AND INFORMATION TRANSMISSION

39

of social determinant and suggest that ‘marker’ can avoid a deterministic baggage. In Sect. 3, I delve into questions of disease causation and explain, at a conceptual level, which notion of causal production should accompany bio- and socio-markers. Finally, in Sect. 4, I discuss good and bad uses of bio- and socio-markers. It is in fact of utmost important that we anticipate potential misuses of the concept. I argue that bio- and sociomarkers are good for explanatory purposes at the individual and aggregate level, and for prediction at the aggregate level; however, they should not be used for predictive purposes at the individual level, as they are likely to introduce bias and discrimination.

2

The Concept of Socio-Marker 2.1

Social Factors as Proximate Causes

Exposure research has made stunning progress in understanding the onset and development of health and disease at the molecular level, by identifying, validating, and use bio-markers. What is now called ‘exposure science’ (Wild 2005, 2012; Vineis, Chadeau-Hyam, et al. 2017; Vineis and Barouki 2022) has developed accredited methodologies to study exposure to numerous environmental hazards at the level of molecules and to trace inside the body how exposure leads eventually to clinical conditions. With bio-markers, we can mark bio-chemical processes of health and disease, gaining very fine-grained understanding of how certain hazards cause disease. Risk factor epidemiology has produced large bodies of evidence that numerous hazards are correlated with health conditions, but it is by pointing at how disease develops that we can get a better grip on causal relations. For instance, nowadays nobody would contest the correlation between smoking and (various types of) cancer, but how exactly does smoking cause cancer, and especially how can smoking cause cancer even years later—these are types of question that molecular epidemiology and bio-markers research has been helping with (Vineis and Russo 2018). However, health and disease are not solely biological phenomena, but they are also social phenomena, with a very broad understanding of ‘social’, as to include social factors proper, political, cultural, demographic, psychological, etc. There are at least two established traditions in the health and in the social sciences that study health and disease in relation to social factors (Russo and Kelly 2023). One strand of the

40

F. RUSSO

literature has been vital to establish, among others, robust correlations between socio-economic inequalities and health inequalities. By and large, however, this literature has established correlations, but has not studied in depth the complex bio-social mechanisms of health and disease and the role social factors play therein. This means, to begin with, that we need to enlarge our concept of disease causation as to include social factors as active or proximate causes of health and disease and not just as remote, classificatory factors. This is needed both for ontological and for normative reasons. On the ontological side, while it is useful to categorize biological factors according to social characteristics, it is a genuine question about the nature of health and disease, and thereby about disease causation whether social factors should be considered proximate (rather than remote causes). If we think that health and disease are phenomena at once biological and social, then our ontology should reflect that. Thus, for instance, obesity has some clearly identified genetic causes. But what is the status of poverty and deprivation? Of course, one may argue that poverty and deprivation cause obesity only in an indirect way, because it is through nutrition and the disfunction of a number of bio-chemical mechanisms that they cause obesity. This line of argument is fallacious for two reasons. The first reason traces back to the ontological claim just made: we should re-learn from sociologists and anthropologists how social processes, forces, conditions, and cultural elements are part and parcel of our lifeworld (see, e.g., ‘Fundamental Cause Theory’ of Link and Phelan (1995) or the discussion and conceptualization of ‘lifeworld’ offered by Russo and Kelly (2023) that, although taking different stances about spelling out the multiple ‘mixed’ mechanisms of health and disease, share the common view of considering social factors as causes proper and not merely classificatory). The second reason, instead, has to do with normative considerations to hold on the view that social factors are proximate causes: if they are proper causes, then we ought to intervene on them in appropriate ways, for instance tackling structural social problems at the societal level, and not just or merely trying to intervene at the level of the individual, specifically, at the bio-chemical, individual level. It is only fair to say that social epidemiology, and part of sociology of health, has produced overwhelming evidence that socio-economic inequalities are correlated with health inequalities. Available quantitative approaches can provide course-grained (but robust) patterns from the social to the biological (Marmot 2005; Mackenbach 2006). Qualitative approaches, instead, can provide very specific individual-level patterns,

SOCIO-MARKERS AND INFORMATION TRANSMISSION

41

with limited generalizability. Another strand of the literature has been fundamental in studying the social dynamics and power structures at place on the social contexts in which health and disease of individuals and of populations happen. By and large, however, this literature has established the working of social factors and social forces at a very high level of abstraction, and in a way that makes it difficult to pinpoint which factors are the ‘actionable’ ones (for a discussion, see Kelly and Russo 2017). 2.2

Social Determinants vs Socio-Markers

The question arises whether we can retain all these valuable elements of the existing approaches to the studies of the social part of health and disease, and attain a level of specificity in the analysis, something analogous to the study of bio-markers in molecular epidemiology. To try and address this question, the concept of socio-marker is worth exploring. To begin with, we can think of health and disease as processes pertaining to individuals or of groups of individuals. We can mark and trace processes of health and disease at the level of the social, just as we mark and trace processes at the level of the biological. To appreciate the importance and relevance of this conceptual shifts towards processes, we need to put the use of (bio- and socio-) markers in epidemiology in the broader context of how epidemiology, across time, has moved from an approach to causation that focused on one-cause–one-effect relations, then to multi-causality (especially thanks to the development and use of ‘causal pies’), and then to reconstructing (causal) processes in which the time component is explicitly considered and modelled. In analogy to bio-markers, the idea of a socio-marker is to identify, validate, and use markers of social processes that intersect, interact, influence, and shape health and disease during the life-course. This would mean a fine-grained tracking of how social markers of early exposure lead to the onset and the development of health and disease at a later stage in life. Although there is an important and established tradition in the health and social sciences that study the links between socio-economic factors and health outcomes and inequalities, this literature is based on the concept of ‘social determinant’ rather than ‘socio-marker’. I take the definition of ‘social determinant’ provided by the World Health Organization (WHO) as a consensus definition, because it is widely adopted in empirical studies and in bioethics discussions. Social determinants are

42

F. RUSSO

[…] non-medical factors that influence health outcomes. They are the conditions in which people are born, grow, work, live, and age, and the wider set of forces and systems shaping the conditions of daily life. These forces and systems include economic policies and systems, development agendas, social norms, social policies and political systems. (WHO 2022)

There is abundant discussion on social determinants and their importance for normative questions, for instance at the level of public health policy, for their potential role in addressing inequalities, etc. (Kelly-Irving et al. 2022). There is, however, much less discussion at the level of epistemology, namely about the baggage of this concept in terms of understanding the very phenomena of health and disease. Reading the WHO definition, it is easy to focus more on the concept of ‘determinant’, giving a connotation that emphasizes more its potential and ability to influence, or determine, later stages of health and disease of individuals than its being related to the social sphere of health and disease. This is somehow reflected not so much in the definition per se, but in the way areas that study the whole intricatedness of the social dimension of health have to constantly establish their legitimacy, against the dominant epidemiological paradigm (Kelly and Russo 2021). On the one hand, semantically, ‘determinant’ points precisely to the power of these factors to ‘influence health outcomes’, and this may divert our attention from the second part of definition, which instead points to the complex social spheres in which these ‘determinants’ act: ‘economic policies and systems, development agendas, social norms, social policies and political systems’. In all likelihood, a deterministic connotation is not intended in the WHO definition of social determinant, and yet the very term ‘determinant’ may carry a deterministic baggage that fires back. The concept of ‘marker’, instead, does not have this baggage. We mark a process, and we trace marks to better understand the complex phenomena of health and disease, making no implicit or explicit reference to how such marker will determine future states of these processes. With the concept of marker, we hopefully skew problems of determinism but, as I argue in Sect. 4, there are other important caveats in using socio(and bio-)markers for predictive purposes. Also, marking a process does not necessarily mean to ‘materially’ intervene and alter a process, but to identify a ‘salient point’ and try to infer future states of the process, which can be done via observational studies and quantitative analysis of data. Reference to ‘key characteristics’ in cancer research is again appropriate

SOCIO-MARKERS AND INFORMATION TRANSMISSION

43

here, because the identification of the intermediate markers of mutation is largely based on observational studies, rather than direct experimental manipulations. A bibliographic search at the time of writing on PubMed for ‘sociomarker’, ‘socio-marker’, and ‘sociomarkers’ returns a handful of publications of which only one explicitly defines the concept. Shin et al. (2018) define socio-markers as ‘measurable indicators of social conditions in which a patient is embedded and is exposed to, being analogous to a bio-marker indicating the severity or presence of some disease state’. They then adopt a machine learning approach to identify asthma patients at risk from hospital record, using both available socio- and bio-markers. This study shows that including socio-markers in the analysis helps a great deal in correctly predicting health outcomes in individual-level variables. The contribution of Ghiara and Russo (2019) does not appear in PubMed database, but it is useful, as it delves into a conceptual discussion of socio-markers. Socio-markers are seen in a complementary way to bio-markers. A socio-marker bears important similarities with ‘social determinant’, in that it is measurable and it refers to social characteristics of individuals (broadly construed as to include economic, social, cultural, or other factors). Socio-markers are primarily individual-level markers, but can be aggregated into group-level measures. Socio-markers are meant to trace, in whole life-course of a individuals, how ‘the social’ has impact on health outcomes and inequalities. Depending on the type of study, with socio-markers we can prospectively predict or retrospectively explain health and social outcomes. Just like bio-markers, socio-markers are not meant to provide full knowledge of complex bio-social mechanisms, but to provide information about ‘key characteristics’ of these mechanisms of health and disease. I will return to this concept in Sect. 3, but it will be useful to anticipate here that, following the most recent developments, ‘mechanism’, as it is mostly used in contemporary philosophy of science, is not a kind of deterministic, Cartesian machine (Glennan and Illari 2018). For my purposes, in particular, it has epistemic, rather than ontic status. There are, however, some important dissimilarities with social determinants. To begin with, typical social science indicators are educational level, occupational class, and income. For socio-markers to be useful, we need to significantly increase the range of relevant indicators, for instance including information about individual behaviour, experience, or interactions, such as exposure to violent behaviour or to certain social structural

44

F. RUSSO

pressures, which of course has important implications for the operationalization and measurement of these markers, and for data collection. It is important to note, however, that in increasing the range of relevant indicators we need to think deeply of which indicators are going to give us the information we are looking for. Thus, for instance, ‘level of education’ needs to be unpacked not so much in terms of determining ‘how much’ an individual is educated in years, months, or days, but rather in terms of, for instance, where or how this education took place. Thus, ‘level of education’ should be used as a construct variable able to disentangle social and cultural aspects: arguably, 10 years of schooling in a deprived neighbourhood in any big city are not equivalent to the same amount of schooling in a rich neighbourhood in the same city. Computational and ‘complexity’ approaches in the health sciences show that we can increase the range and granularity of social variables (for empirical studies applying a complexity approach, see, e.g., Crielaard et al. 2021, Merabet et al. 2022). The granularity, especially, does not refer merely to the possibility of measuring a same variable in a more precise way quantitatively (e.g. months rather than years of schooling), but qualitatively, namely by spelling out the various socio-eco-demo-cultural-political processes that are at work in the proxy ‘level of education’. This does come with challenges and costs (for instance for data collection), but it is not the main barrier. Instead, the most important change is in the use and interpretation of data as markers of social processes. This is a good moment to return to the very concept of marker, and how it differs from a social determinant. A socio-marker indicates, or marks, some state of the process, rather than determining, future stages of the process. As mentioned earlier, it is important not to conflate causes and markers, as markers may be causes, but very often are not. Yet, the process they mark is causal. The importance of introducing the concept of socio-marker is double. On the one hand, it allows us to have an analysis of social factors that is more specific than the social determinants one. On the other, it helps giving social factors a prominent causal role, because socio-markers have to be understood as markers of processes that are proximate, not just remote or, even worse, as mere classificatory factors. I have identified some similarities and dissimilarities between bio- and socio-markers, but there is also another point that deserves discussion. We tend to think that bio-markers are very specific in identifying precise characteristics of the targeted process at the biological level. Thus, for instance, once we establish that DNA methylation is a relevant process

SOCIO-MARKERS AND INFORMATION TRANSMISSION

45

to understand the causal role of environmental stressors, measurement of methylation levels in tissues can be very precise, and it is the precision of such measurements across groups that will allow us to reconstruct (aspects of) the carcinogenic or other pathogenic processes, via the identification of ‘salient moments’ or ‘key characteristics’. The question arises whether, and to what extent, socio-markers can attain the same level of specificity. We need to avoid the temptation of using ‘quantifiable’ precision as the divide between natural (including health) sciences and the social sciences, the former being ‘exact’, but not the latter. The precision of the measurement of certain bio-chemical processes, such as DNA methylation, rests on the assumption of biological homogeneity across individuals. And while this assumption is questioned already from within the health sciences (think for instances at the studies on the microbiome), it should be safe to assume some level of biological homogeneity among individuals—or at least this should be a reasonable working hypothesis. At the level of the social, however, we do know that individuals follow unique paths in the life-course, responding to similar situations in very different ways. And yet, the social sciences are also able to identify generic patterns that individuals follow, and the argument for socio-markers rests on the premise that the robust and stable correlations identified by the social determinants approach can be spelled out in terms of more precise processes, for instance disentangling psychological processes proper, social pressures (likely highly cultural-dependent), political conditions, etc. This level of specificity is clearly not necessarily quantifiable in the same way as DNA methylation is, but the qualitative level of specificity is instead comparable. Here, quantification refers to the exact measurement of levels of DNA methylation, which can be attained only thanks to the sophisticated instrumentation used in molecular epidemiology (see Russo 2022). The qualitative level of specificity, instead, refers to the role DNA methylation has in the broader mechanism of cancer development, and spelling out roles and functions is core business of social science approaches too. These considerations, to be sure, are not to re-instate a division between the natural science and the social science, where the former are exact, dealing with quantifiable variables, and the latter remain after all inexact, dealing with qualitative variables. If anything, these considerations challenge this division and call for more qualitative reasoning to support numbers, and for care in equating the possibility of quantifying

46

F. RUSSO

with the ability of generalizing. The social sciences, although not quantifiable in the same way as the natural, bio-chemical sciences, can attain generalizations and also be very specific about individual paths.

3 How to Trace Information Transmission with Bio- and Socio-Markers 3.1

Causal Production in the Processes of Health and Disease

The introduction of the concept of marker is accompanied with another important conceptual change. Instead of talking about causes of health and disease as such, biological and social causes are part of causal processes. To be able to work with the concept of socio-marker (and with biomarker), we need to adopt a specific approach to disease causation (information transmission) and to employ mixed and multiple methodologies for the study of health and disease. In this section, I develop in some detail the idea of causal production as information transmission and only make a quick reference to the need of pluralistic methodological approach, which I leave for future work. In turn, causal production as information transmission is very akin to thinking of health and disease along the lines of processes of pathogenesis and salutogenesis . Interestingly, ‘pathogenesis’, in the scientific and philosophical literature alike, is often referred to as a process leading to some specific disease. Although the study of pathogenic processes is largely based on the conceptual tenets of risk factor epidemiology and ‘entity-based’ pathophysiology (Carter 2003), some recent contributions have considered more specifically the prospects of a process perspective in medicine in particular and in the sciences more generally (see, respectively, Dammann 2020 and Russo 2022), and the potential of salutogenesis for developing a science of population health (Valles 2019). Moreover, the concept of salutogenesis may help us give due importance to social factors and a fortiori to socio-markers. The concept of salutogenesis (Antonovsky 1965) is quite close to a processual understanding of health and disease, and therefore to a lifecourse approach in which socio-markers may play an important role. To appreciate the importance and relevance of salutogenesis, it will be useful to start from the widely accepted definition of health of the World Health

SOCIO-MARKERS AND INFORMATION TRANSMISSION

47

Organisation1 that points not just to the absence of disease but also to the well-being of individuals, in turn not reducible to biology only. The salutogenic approach proposed by Antonovsky reframed the usual biomedical question of what causes disease in terms of what, instead, leads some people rather than others to stay healthy, despite enduring stressful circumstances (Eriksson and Lindstrom 2007; Burns 2014; Mittelmark and Bauer 2017). ‘Sense of coherence’ is the concept at the basis of the salutogenic model; this includes life experiences across social and biological factors, and starts early in life. To understand its origins and development, we need to delve into the cultural and historical contexts in which individuals live, and the kind of coping mechanisms they develop to face all sorts of situation, and specifically the one that generates psychosocial stressors and other resistance resources. It is important to realize that not everyone will react to situations that are similar in the same way, and for this reason, we need as well to consider ‘intra-person and extra-person differentiation’, which introduces a community- and population-level perspective to health, next to an individual-level one. In the reconstruction of Mittelmark and Bauer (2017, 11), Antonosky developed his salutogenic approach as an explicit reaction to a form of reductionistic thinking, typical of the Western medical approach, and according to which the human body is a beautiful system, machine-like, that is put under attack by pathogens, viz. the causes of disease. The salutogenic approach thus challenges a sharp dichotomy between healthy and pathological, suggesting that there is instead a continuum, and that we need to get a better grip on the factors that promote health. Through the articulation of the concept of ‘sense of coherence’, these factors are not just to be located at the level of (individual) biology, but also at the level of the social and at the individual, community, and population level (Burns 2014). I return in Sect. 3.2 to the usefulness of the salutogenic approach, in relation to socio-markers. I refer here to salutogenesis to introduce the concept of ‘process’ and of ‘production’. This invites for a conceptual change—from studying the causes of health and disease to studying the processes of health and disease. In turn, this conceptual shift does not mean that just increasing the

1 See https://www.who.int/about/governance/constitution, December 2022.

accessed

on

8th

48

F. RUSSO

granularity and precision of quantitative models will do. From a methodological perspective, instead, they need mixed- (or multi-)methodologies that combine qualitative and quantitative approaches, as well as multilevel ones as for instance the ones employed in demography or education research, that are able to combine, in a single statistical framework, aggregate- and individual-level variables. But there is more. In both the discussion of Dammann and of Valles, there is an implicit discontent with the risk factor approach, because alone, it cannot tell us enough about the processes of health and disease. However, while Dammann (2020) provides a pretty convincing story for why focusing on processes complements the standard story about disease causation framed in terms of individual factors and mechanisms, Valles (2019) also makes the important point that the relevant processes are not just those of disease but those of health. It is by focusing on processes of health (salutogenesis) that we can aim to improve on the health of populations, and arguably especially about those populations that are more needing. While it has been established that, in general, the wealthier are also healthier, it is by studying the specifics of this generic claim that we can hope to drive change. Valles appeals to the ‘Fundamental Cause Theory’ as an important approach to drive change (Link and Phelan 1995; Phelan and Link 2005; Phelan et al. 2010), and I take the programmatic introduction of ‘socio-markers’ as a step in the same direction, although the accounts may give different emphasis to (mixed) mechanisms of health and disease. Yet, the introduction of ‘socio-markers’ is not unproblematic. Giroux et al. (2021) (and also Giroux, this volume) express the worry that socio-markers will reinforce the ‘mechanistic model’ of bio-markers and express concerns that this may lead to forms of reductionism. The worry is sound and real, I think. However, studying socio-markers should be methodologically placed in a broader discussion of how to ‘operationalize’ the social, not to reduce it to the biological. It is by focusing on questions of operationalization and measurement in pluralistic methodological perspective that we can make socio-markers epistemically more ‘usable’, especially for policy purposes. This is of course a difficult and lengthy exercise, but one that has been initiated. For instance, Russo and Kelly (2023) discuss the use of both qualitative and quantitative approaches to better understand the ‘lifeworld’ of health and disease; Vineis (2020) and Vineis and Barouki (2022) try to integrate the sociological framework of Pierre Bourdieu into the life-course approach;

SOCIO-MARKERS AND INFORMATION TRANSMISSION

49

and Kelly-Irving and Delpierre (2017, 2021) discuss the concept of embodiment to understand social-to-biological processes. This is clearly an important line of research, and by developing it further, we can hopefully appropriately address, and defuse, the charge of reductionism. I will not pursue this line on this occasion and instead concentrate on questions related to the concept of marker and of causal production that should accompany it. 3.2

The Concept of Information Transmission

We face here a conceptual difficulty, as shifting towards processual understandings of health and disease opens the floor for questions about the meaning of causation. If the focus is not the kind of usual causal framework of ‘risk factors’, what kind of notion of causation may be of help? The focus on processes of health and disease can be categorized, following Illari and Russo (2014), as accounts of (causal) production; in turn, causal production is classically considered as primarily question about (causal) metaphysics. This means providing a philosophical account about what the linking between causes and effects is, which is ultimately a question about causal metaphysics (Illari 2011; Illari and Russo 2016b). It is, however, important to note that questions about causal metaphysics, although distinct, are not entirely independent of questions of causal epistemology and/or methodology (Illari and Russo 2014). So, what kind of productive causality is at work in biomedical contexts? I have presented and discussed the concept of productive causality, especially in the context of molecular epidemiology, in some joint publications, and will summarize the main arguments here (Illari and Russo 2016a, b; Russo and Vineis 2016; Vineis, Illari, and Russo 2017; Vineis and Russo 2018). In the philosophy of causality, the question of causal production has received a range of answers, appealing to concepts of powers, capacities, or dispositions; this strand of the literature has been, admittedly, quite ‘metaphysically-oriented’, and at times in a way that is rather disconnected from the study of the practice of science. Other approaches, more ‘science-oriented’, cashed out causal production in terms of mechanisms. Arguments for causal productions are also given in terms of the notion of process. I will not repeat here arguments given in previous publications, in which it is explained why all these approaches are useful but also of limited applicability. I instead directly introduce

50

F. RUSSO

the concept of information transmission. This concept builds on the socalled Salmon-Dowe approach, which hinges on the idea of ‘process’, and broadens it significantly. The goal of ‘information transmission’ is to provide a general account of causal production, one that is able to work across the micro- and macro-world, and across inhomogeneous such as social and biological factors. For these reasons, information transmission helps with offering a suitable concept of causal production for bio-markers and by extension for socio-markers too. Information transmission, as the account of causal production that cashes out the idea that bio- and sociomarkers, helps us trace the causal role of biological and socio-economic causes, and also how these two may interact. To appreciate the potential of information transmission, it is important to return to the idea that markers (bio or social) do not provide the full description of bio-social mechanisms of health and disease, but instead give key characteristics, or salient moments, in these mechanisms. I have primarily used, so far, the concept of process, and here I am introducing the concept of mechanism. Let me explain why there is no contradiction in this shift of terminology. Both ‘mechanism’ and ‘process’ are notions thoroughly discussed in philosophy of science, and of medicine for the matter. I see no inherent tension in switching from one to the other, but it is important to emphasize and explain when one is better suited than another. When using ‘mechanisms’ of health and disease, I give emphasis to explanation, and especially to how the specific arrangements and organization of different factors within the mechanism help explain the phenomenon under scrutiny. In so doing, I work with a minimal definition of mechanism, as is defended by Glennan et al. (2022, 145): The most fruitful way to define mechanisms is that a mechanism for a phenomenon consists of entities (or parts) whose activities and interactions are organized so as to be responsible for the phenomenon’.

A minimal definition such as the one above allows for mixed mechanisms, i.e. those in which both biological and social factors play a role (Kelly et al. 2014; Kelly and Russo 2017; Russo and Kelly 2023). The use I make of ‘mechanism’ in this context is primarily epistemic. Instead, when using ‘process’ of health and disease, I give emphasis to metaphysical aspects. A processual understanding of health and disease, in particular, ties well with a life-course approach, in which we try to make sense of health

SOCIO-MARKERS AND INFORMATION TRANSMISSION

51

and disease in the long term, starting in utero, and continuing through the whole course of life. But in reconstructing the process of health and disease of an individual (or group of individuals), we do not need to fill all the spaces between two points with a continuous line. It suffices to be able to identify key moments, with an incomplete dotted or dashed line, to remain within the metaphor. For instance, measuring exposure to violent behaviour within household during childhood is meant to identify a salient ‘moment’ in early life, likely to have an impact on later stages of health and of social conditions. How exactly does ‘violent behaviour in childhood’ cause health and social outcomes at later stages in life is something that we seek to explain via a social science approach (qualitative and quantitative), by understanding and reconstructing individual experiences when and if needed, and via a biomedical science approach proper, by understanding how an experience such as violent behaviour in childhood ‘gets under the skin’ and becomes in a sense detectable at the level of biochemistry of health and disease. This is the sense in which socio-markers proper and bio-markers of social processes are complementary. Having clarified the use of ‘mechanism’ and of ‘process’, let me return to ‘information transmission’. The account of causal production as information transmission has been developed in previous work, as mentioned earlier. Details of the account, including its roots in Salmon’s original account based on processes and its place in the ‘causal mosaic’, are discussed in these publications. I will highlight here its main features. To begin with, it is useful to recover a concept of process that is more general than a physics process. This was initially given by Salmon, who in turn relied on the concept of ‘world-line’ developed by Reichenbach. The process account later developed by Salmon (and then in the debate with Dowe too) was, however, tailored to physics, identifying conserved quantities such as energy or momentum as the ‘markers’ to distinguish between causal and non-causal processes. The question of causal production can then be posed as to how, by marking a process of some kind, a mark is to be found at later points of the same process. Thus, for instance, two billiard balls colliding are causal processes that intersect at some point, and because of this interaction the billiard balls modify their velocity or direction—measurements related to velocity or direction are markers of the causal process. Instead, the shadows of two airplanes crossing on the ground are not modified after the interaction—here there is no relevant marker of the interaction between the airplanes.

52

F. RUSSO

How can we make the concept of process more general than this? This is where the concept of information kicks in. The first informational account of causal production was that of Collier, who interpreted processes and the transmission of marks in informational terms. The point of an informational account of causal production is not so much to reduce or re-describe any reality in 0s and 1s, but to conceptualize causal production differently; in this case, the transmission of information is what cashes out the productive aspect of causality, and we can use socio-markers to intercept the transmission of information, from exposure to, say, adverse childhood experiences, until disease develops at later moments in life. With a general account of causal process, we can make sense of causal production across levels: micro–macro, biological-social. We can give metaphysical ground to claims that social factors are causes of health and disease (not just classificatory factors) and to claims that environmental hazards (macro) have causal effects at the level of molecules of individuals (micro). It is a thin metaphysics, in that it does not presuppose any materialist account of causation, and especially a reduction to physics or to bio-chemistry. This is a good moment to return to the salutogenic approach. As mentioned in Sect. 3.1, the ‘sense of coherence’ is key to understand how some people (but not others) remain healthy despite being exposed to similar stressors. A social determinant approach establishes that poverty is linked to poor health; a reductionistic, biomedical approach would explain away this link by referring to, e.g., poor diet. A salutogenic approach, however, would require us to provide thick descriptions of how individuals respond to stressful situations, to lack of financial resources, poor living conditions, etc. To get an empirical grip on these thick descriptions, socio-markers may help, because we are forced to ‘unpack’ the course-grained determinant ‘poverty’ into more informative markers of the numerous processes at work. Bio- and socio-markers help us trace relevant causal processes that happen across the biological and social realms, and across the individual and group level. As Vineis (2020) says, the key question is to connect, in appropriate ways, the ‘zoe’ (biological life, including biological mechanisms and outcomes) and the ‘bios’ (biographical life, including the social environment and social dynamics broadly construed). In a lifecourse perspective, a socio-marker such as ‘adverse child experience’ is connected to a number of biological ‘fingerprints’ of adverse conditions, at later stages in life. In this approach, measuring early childhood adverse experience using individual-level variables can help us understand more

SOCIO-MARKERS AND INFORMATION TRANSMISSION

53

general causal processes which a number of individuals are part of, and that show other measurable effects in terms of health outcomes later in life, possibly aggregating results at group level. Introducing socio-markers comes with important considerations, methodological and conceptual. At the methodological level, using socio-markers requires working with variables measured at the individual-level; we may need to generate and collect new data sets for this purpose, in case existing databases with measurements of social variables are not specific enough. At the conceptual level, using socio-markers implies embracing the view that they are not mere classificatory factors; instead, socio-markers mark relevant points or characteristics of the (life-long) causal process(es) that lead to health and disease, often much later in time. The use of socio-markers is thus not just in line, but may further help developing empirical studies of salutogenesis, by appropriately marking the various processes that contribute to the ‘sense of coherence’, and that are located at the level of individual experience as well as at the level of environment, group dynamics, or cultural factors. So far, I tried to provide further detail on the concept of socio-marker and to recap its conceptual underpinning in an account of causal production in terms of information transmission. I discussed conceptual and methodological aspects of bio-markers: we need them both—because of the bio-social nature of phenomena of health and disease, and because we want to say that social factors are causes. With socio-markers we can mark and trace relevant processes in which social factors play a role, and trace this information flow until effects of social causes are ‘visible’ at either the social or biological level, or both. The job of epistemology is, however, not complete; much more would need to be discussed, notably related to methodological protocols for data generation, collection, and analysis. In the final section, however, I focus on normative implications related to introducing the concept of socio-marker.

4

Using Bio- and Socio-Markers

Concepts and methods are not neutral ‘instruments’. There is abundant literature on science and values, documenting and showing how values do play a role in research (Carrier 2013; Douglas 2009; Elliott 2017; Gonzalez 2013; Kincaid et al. 2007). This is how science is laden with values. But we may want likewise to ask the question of which values we promote, but using certain concepts and methods (Russo 2021). Bio- and

54

F. RUSSO

socio-markers are no exception and, far from being ‘value-free’, they can be promotors of values we would rather not promote. In this section, I discuss potential and actual implications, as well as intended and nonintended uses of bio- and socio-markers. The argument here makes a normative twist, about how to use and not to use bio- and socio-markers. Briefly put, using bio- and socio-markers should contribute to: • Understanding the bio-social mechanisms of health and disease; • Predicting population-level health and social outcomes, from early events in life or from specific biological conditions; • Designing public health interventions and individual treatment, taking into account these markers. Let me elaborate on these in more detail. In the previous sections, I presented exposure research as the latest frontier of epidemiology, and as an area that has greatly advanced our understanding of health and disease, at the bio-chemical level. I also introduced the concept of socio-marker to suggest that the lifecourse approach could be thereby complemented. Bio- and socio-markers thus positively contribute to understanding the mechanisms of health and disease, especially looking at their interactions, and thus making a bio-psychosocial model operational. The potential of using bio- and socio-markers, however, is not confined to (retrospectively) explain the mixed mechanisms of health and disease, but also to predict health and social outcomes, at the population level. Exposure research (complemented with a socio-marker component) retains the hallmark of epidemiology, namely its population-level character. It is important that we attempt to project the status of social and health outcomes, because it is also on the basis of these projections that we can establish priorities for further research and for policy making. The population-level character of prediction using bio- and sociomarkers is closely connected to the design of public health interventions. It is worth noting that while bio-markers may not be actionable in a public health perspective, socio-markers may instead be much more informative about the specific joints to target at the level of public health. Evidence coming from bio- and socio-marker studies can in fact provide evidence to support specific interventions (Kelly-Irving et al. 2022).

SOCIO-MARKERS AND INFORMATION TRANSMISSION

55

Concerning individual treatment, I will not comment directly on the use of bio-markers, as these are already routinely used, for instance to tailor cancer therapies. Using socio-markers for individual treatment instead remains at this very moment quite speculative and would require much more synergies between medical practitioners, psychologists, and numerous other health professional figures such as nurses and midwives. In general, using bio- and socio-markers for inferences at the aggregate or population level should be pretty safe, unless of course an intervention is intended to discriminate or marginalize a whole group, identifiable via a specific bio- or socio-marker. Instead, we should refrain from any individual predictive use for any purpose that is outside the remit of explanation (e.g. diagnosis and prognosis) and treatment planning. Let me explain. Individual predictive use of bio- and socio-markers should not be used as an indicator of future states in life of a real person. There is in fact a difference between working with individual-level variables and making a prediction in the single case, for a real individual. This is perhaps an advantage that social determinants carry over sociomarkers. Given how social determinants are defined and measured, they cannot really pick on individual (social) characteristics. Instead, just as we do with bio-markers, with socio-markers we intend to measure an individual-level characteristics, for instance a specific type of exposure to adverse experience in childhood, or exposure to poor living conditions in a given neighbourhood, or lack of social support in a poor education environment. The use of individual-level variables gives the wrong impression that a prediction in the single case, for a real person, is automatically legitimate, but it is not. Individual-level variables are generic, after all. The whole point of gathering information about social processes at this level of specificity is to understand (see good uses above), in a lifecourse approach, how early exposure to certain social conditions is part of processes of pathogenesis or salutogenesis. But this individual-level information should not be used in any way that could discriminate individuals in the future. Misuses of bio-markers, leading to discriminatory practices, are already discussed and documented (Pollitz et al. 2007; Davis 2010; Lakhan et al. 2010; Arias et al. 2018). The way in which, say, health insurances use bio-markers is a gross and dangerous mis-interpretation, one that gives bio-markers (and potentially socio-markers too) deterministic flavour. In this way, bio- and socio-markers can be easily used to justify and validate discriminatory practices. We should be careful not to repeat the

56

F. RUSSO

same mistakes, if socio-markers are to be used in empirical research in the future. If there is any good prospect in using the concept of socio-marker, its scope and applicability should be clarified from the start. Providing very specific provisions about use of bio- and socio-markers is a difficult matter that also involves delicate and subtle legal considerations that are likely to depend on local legislation too. However, what the scientific community can do is to keep reflecting explicitly, in published work, about what can legitimately be inferred from any given study that uses bio- and socio-markers. This should go beyond usual caveats about inferring causation from correlation and should instead contain clearer stances about the elements indicated above: explanatory, predictive, evidential, and at the individual or aggregate level.

5

Conclusion

Health and disease are complex bio-social phenomena. This is no breaking news, and in fact, the idea has a lengthy pedigree (Antonovsky 1965; Engel 1980). And yet, the stunning advancement of molecular biology and molecular epidemiology has contributed to shift emphasis to the biochemical processes of health and disease, with a relative neglect of the social part. How can we restore a balance methodologically? In this paper, I suggested that introducing the concept of socio-marker, next to the one of bio-marker, can help studying the bio-social processes of health and disease, at the same level of specificity, thus not confining social factors to mere classificatory devices and avoiding the possible ‘deterministic’ baggage of the social determinants literature. If the concept of socio-marker holds a promise, we need to collect data and generate evidence accordingly. Most importantly, however, the promise of bio- and socio-marker should not make us blind towards normative implications. Specifically, we should refrain from using bio- and socio-markers for individual predictive purposes, because this use is likely to introduce bias and discrimination, whether intentionally or not. Their usefulness is instead in aggregate prediction and in explanation (aggregate- or individual-level), and this good use holds, hopefully, a promise to take the study of social factors to the next stage. Acknowledgments This chapter elaborates upon published work with Virginia Ghiara. I am hugely indebted to Élodie Giroux for her encouragement to develop

SOCIO-MARKERS AND INFORMATION TRANSMISSION

57

the concept of ‘socio-markers’ further. I hope that the way I elaborated on the original formulation of the concept makes a valuable contribution to the literature. All comments received during the workshop “Epistemological and practical issues of integrative approaches in environmental health” held in February 2021 have been invaluable. I am also very grateful to Paolo Vineis for the opportunity to present a draft of this paper at one of the meetings with his group in June 2022, where I received very many useful suggestions to clarify several passages. Comments from anonymous reviewers are also gratefully acknowledged. Any error or inaccuracies remain of course mine.

References Antonovsky, Aaron. 1965. ‘Social Class, Life Expectancy and Overall Mortality’. Milbank Memorial Fund Quarterly 45 (2): 31–73. https://doi.org/10. 2307/3348839. Arias, Jalayne J., Ana M. Tyler, Benjamin J. Oster, and Jason Karlawish. 2018. ‘The Proactive Patient: Long-Term Care Insurance Discrimination Risks of Alzheimer’s Disease Biomarkers’. Journal of Law, Medicine & Ethics 46 (2): 485–98. https://doi.org/10.1177/1073110518782955. Berger, Eloïse, Raphaële Castagné, Marc Chadeau-Hyam, Murielle Bochud, Angelo d’Errico, Martina Gandini, Maryam Karimi, et al. 2019. ‘MultiCohort Study Identifies Social Determinants of Systemic Inflammation Over the Life Course’. Nature Communications 10 (1): 773. https://doi.org/10. 1038/s41467-019-08732-x. Biomarkers Definitions Working Group. 2001. ‘Biomarkers and Surrogate Endpoints: Preferred Definitions and Conceptual Framework’. Clinical Pharmacology & Therapeutics 69 (3): 89–95. https://doi.org/10.1067/mcp. 2001.113989. Blane, David, Michelle Kelly-Irving, Angelo d’Errico, Melanie Bartley, and Scott Montgomery. 2013. ‘Social-Biological Transitions: How Does the Social Become Biological?’ Longitudinal and Life Course Studies 4 (2). https:// doi.org/10.14301/Llcs.V4i2.236. Burns, Henry. 2014. ‘What Causes Health?’ The Journal of the Royal College of Physicians of Edinburgh 44 (2): 103–5. https://doi.org/10.4997/JRCPE. 2014.202. Califf, Robert M. 2018. ‘Biomarker Definitions and Their Applications’. Experimental Biology and Medicine 243 (3): 213–21. https://doi.org/10.1177/ 1535370217750088. Carrier, Martin. 2013. ‘Values and Objectivity in Science: Value-Ladenness, Pluralism and the Epistemic Attitude’. Science & Education 22 (10): 2547– 68. https://doi.org/10.1007/s11191-012-9481-5.

58

F. RUSSO

Carter, K. Codell. 2003. The Rise of Causal Concepts of Disease: Case Histories. Burlington, VT: Ashgate. Crielaard, Loes, Mary Nicolaou, Alexia Sawyer, Rick Quax, and Karien Stronks. 2021. ‘Understanding the Impact of Exposure to Adverse Socioeconomic Conditions on Chronic Stress from a Complexity Science Perspective’. BMC Medicine 19 (1): 242. https://doi.org/10.1186/s12916-021-02106-1. Dammann, Olaf. 2020. Etiological Explanations: Illness Causation Theory. Boca Raton: CRC Press. Davis, John K. 2010. ‘Justice, Insurance, and Biomarkers of Aging’. Experimental Gerontology 45 (10): 814–18. https://doi.org/10.1016/j.exger.2010. 02.004. Douglas, Heather E. 2009. Science, Policy, and the Value-Free Ideal. Pittsburgh, PA: University of Pittsburgh Press. Elliott, Kevin Christopher. 2017. A Tapestry of Values: An Introduction to Values in Science. New York, NY: Oxford University Press. Engel, George L. 1980. ‘The Clinical Application of the Biopsychosocial Model’. American Journal of Psychiatry 137 (5): 535–44. https://doi.org/10.1176/ ajp.137.5.535. Eriksson, Morika, and Bengt Lindstrom. 2007. ‘Antonovsky’s Sense of Coherence Scale and Its Relation with Quality of Life: A Systematic Review’. Journal of Epidemiology & Community Health 61 (11): 938–44. https://doi.org/10. 1136/jech.2006.056028. Errico d’, Angelo, Fulvio Ricceri, Silvia Stringhini, Cristian Carmeli, Mika Kivimaki, Mel Bartley, Cathal McCrory, et al. 2017. ‘Socioeconomic Indicators in Epidemiologic Research: A Practical Example from the LIFEPATH Study’. PLOS ONE 12 (5): e0178071. https://doi.org/10.1371/journal.pone.017 8071. Eybpoosh, Sana, Ali Akbar Haghdoost, Ehsan Mostafavi, Abbas Bahrampour, Kayhan Azadmanesh, and Farzaneh Zolala. 2017. ‘Molecular Epidemiology of Infectious Diseases’. Electronic Physician 9 (8): 5149–58. https://doi.org/ 10.19082/5149. Ghiara, Virginia, and Federica Russo. 2019. ‘Reconstructing the Mixed Mechanisms of Health: The Role of Bio- and Sociomarkers’. Longitudinal and Life Course Studies 10 (1): 7–25. https://doi.org/10.1332/175795919X15468 755933353. Giroux, Élodie, Yohan Fayet, and Thibaut Serviant-Fine. 2021. ‘L’Exposome: Tensions Entre Holisme et Réductionnisme’. Médecine/Sciences 37 (8–9): 774–78. https://doi.org/10.1051/medsci/2021092. Glennan, Stuart, and Phyllis Illari, eds. 2018. The Routledge Handbook of Mechanisms and Mechanical Philosophy. Routledge.

SOCIO-MARKERS AND INFORMATION TRANSMISSION

59

Glennan, Stuart, Phyllis Illari, and Erik Weber. 2022. ‘Six Theses on Mechanisms and Mechanistic Science’. Journal for General Philosophy of Science 53 (2): 143–61. https://doi.org/10.1007/s10838-021-09587-x. Gonzalez, Wenceslao J. 2013. ‘Value Ladenness and the Value-Free Ideal in Scientific Research’. In Handbook of the Philosophical Foundations of Business Ethics, edited by Christoph Luetge, Dordrecht: Springer Netherlands, 1503–21. https://doi.org/10.1007/978-94-007-1494-6_78. Honardoost, Maryam, Azam Rajabpour, and Ladan Vakil. 2018. ‘Molecular Epidemiology; New but Impressive’. Medical Journal of the Islamic Republic of Iran 32 (1): 312–16. https://doi.org/10.14196/mjiri.32.53. Illari, Phyllis McKay. 2011. ‘Why Theories of Causality Need Production: An Information-Transmission Account’. Philosophy and Technology 24 (2): 95– 114. https://doi.org/10.1007/s13347-010-003-3. Illari, Phyllis McKay, and Federica Russo. 2014. Causality: Philosophical Theory Meets Scientific Practice. Oxford, UK: Oxford University Press. Illari, Phyllis, and Federica Russo. 2016a. ‘Causality and Information’. In The Routledge Handbook of Philosophy of Information, edited by Luciano Floridi, Routledge. 235–48. ———. 2016b. ‘Information Channels and Biomarkers of Disease’. Topoi 35 (1): 175–90. https://doi.org/10.1007/s11245-013-9228-1. International Programme on Chemical Safety, ed. 2001. Biomarkers in Risk Assessment: Validity and Validation. Environmental Health Criteria 222. Geneva: World Health Organization. Kelly, Michael P., Rachel S. Kelly, and Federica Russo. 2014. ‘The Integration of Social, Behavioral, and Biological Mechanisms in Models of Pathogenesis’. Perspectives in Biology and Medicine 57 (3): 308–28. https://doi.org/ 10.1353/pbm.2014.0026. Kelly, Michael P., and Federica Russo. 2017. ‘Causal Narratives in Public Health: The Difference Between Mechanisms of Aetiology and Mechanisms of Prevention in Non-Communicable Diseases’. Sociology of Health & Illness 40 (1): 82–99. https://doi.org/10.1111/1467-9566.12621. ———. 2021. ‘The Epistemic Values at the Basis of Epidemiology and Public Health’. MEFISTO. Journal of Medicine, Philosophy, and History 5 (1): 1–26. https://journal.edizioniets.eu/index.php/mefisto/article/view/342. Kelly-Irving, Michelle, and Cyrille Delpierre. 2017. ‘The Embodiment Dynamic Over the Life Course: A Case for Examining Cancer Aetiology’. In The Palgrave Handbook of Biology and Society, edited by C. Meloni, J. Cromby, D. Fitzgerald, and S. Lloyd, London: Palgrave Macmillan. 519–40. Kelly-Irving, Michelle, and Cyrille Delpierre. 2021. ‘Framework for Understanding Health Inequalities Over the Life Course: The Embodiment Dynamic and Biological Mechanisms of Exogenous and Endogenous Origin’.

60

F. RUSSO

Journal of Epidemiology and Community Health 75 (12): 1181–86. https:// doi.org/10.1136/jech-2021-216430. Kelly-Irving, Michelle, William Patrick Ball, Clare Bambra, Cyrille Delpierre, Ruth Dundas, Julia Lynch, Gerry McCartney, and Katherine Smith. 2022? ‘Falling Down the Rabbit Hole? Methodological, Conceptual and Policy Issues in Current Health Inequalities Research’. Critical Public Health. 33 (1) 37–47. https://doi.org/10.1080/09581596.2022.2036701. Kelly-Irving, Michelle, Silke Tophoven, and David Blane. 2015. ‘Life Course Research: New Opportunities for Establishing Social and Biological Plausibility’. International Journal of Public Health 60 (6): 629–30. https://doi.org/ 10.1007/s00038-015-0688-5. Kincaid, Harold, John Dupré, and Alison Wylie, eds. 2007. Value-Free Science? Ideals and Illusions. Oxford, New York: Oxford University Press. Lakhan, Shaheen E., Karen Vieira, and Elissa Hamlat. 2010. ‘Biomarkers in Psychiatry: Drawbacks and Potential for Misuse’. International Archives of Medicine 3 (1): 1. https://doi.org/10.1186/1755-7682-3-1. Link, Bruce G., and Jo C. Phelan. 1995. ‘Social Conditions as Fundamental Causes of Disease’. Journal of Health and Social Behavior Spec No: 80–94. https://pubmed.ncbi.nlm.nih.gov/7560851/. Mackenbach Johan P. 2006. ‘Socio-Economic Inequalities in Health in Western Europe: From Description to Explanation to Intervention’. In Social Inequalities in Health: New Evidence and Policy Implications, edited by Johannes Siegrist and Michael Marmot. Oxford University Press. 223-250. https:// doi.org/10.1093/acprof:oso/9780198568162.003.00010. Marmot, Michael. 2005. ‘Social Determinants of Health Inequalities’. The Lancet 365 (9464): 1099–1104. https://doi.org/10.1016/S0140-6736(05)711 46-6. Merabet, Nadège, Paul J. Lucassen, Loes Crielaard, Karien Stronks, Rick Quax, Peter M. A. Sloot, Susanne E. la Fleur, and Mary Nicolaou. 2022. ‘How Exposure to Chronic Stress Contributes to the Development of Type 2 Diabetes: A Complexity Science Approach’. Frontiers in Neuroendocrinology 65 (April): 100972. https://doi.org/10.1016/j.yfrne.2021.100972. Mittelmark, Maurice B., and Georg F. Bauer. 2017. ‘The Meanings of Salutogenesis’. In The Handbook of Salutogenesis, edited by Maurice B. Mittelmark, Shifra Sagy, Monica Eriksson, Georg F. Bauer, Jürgen M. Pelikan, Bengt Lindström, and Geir Arild Espnes, Cham: Springer International Publishing. 7–13. https://doi.org/10.1007/978-3-319-04600-6_2. NCI Dictionaries. 2022. ‘Biomarker’. In Dictionary of Cancer Terms. https:// www.cancer.gov/publications/dictionaries/cancer-terms/def/biomarker. Phelan, Jo C., and Bruce G. Link. 2005. ‘Controlling Disease and Creating Disparities: A Fundamental Cause Perspective’. The Journals of Gerontology:

SOCIO-MARKERS AND INFORMATION TRANSMISSION

61

Series B 60 (Special_Issue_2): S27–S33. https://doi.org/10.1093/geronb/ 60.Special_Issue_2.S27. Phelan, Jo C., Bruce G. Link, and Parisa Tehranifar. 2010. ‘Social Conditions as Fundamental Causes of Health Inequalities: Theory, Evidence, and Policy Implications’. Journal of Health and Social Behavior 51 (1_suppl): S28–40. https://doi.org/10.1177/0022146510383498. Pollitz, Karen, Beth N. Peshkin, Eliza Bangit, and Kevin Lucia. 2007. ‘Genetic Discrimination in Health Insurance: Current Legal Protections and Industry Practices’. INQUIRY: The Journal of Health Care Organization, Provision, and Financing 44 (3): 350–68. https://doi.org/10.5034/inquiryjrnl_44. 3.350. Rothman, Nathaniel, Pierre Hainaut, Paul Shulte, and Frederica Perera, eds. 2011. Molecular Epidemiology: Principles and Practices. IARC Scientific Publications, no. 163. Lyon, France: International Agency for Research on Cancer, World Health Organization. Russo, Federica. 2021. ‘Value-Promoting Concepts in the Health Sciences and Public Health’. Preprint. PhilSci Archive. http://philsci-archive.pitt.edu/id/ eprint/19287. ———. 2022. Techno-Scientific Practices: An Informational Approach. Rowman and Littlefield International. Russo, Federica, and Michael P. Kelly. 2023. ‘The ‘lifeworld’ of Health and Disease and the Design of Public Health Interventions’. Longitudinal and Life Course Studies. In press. Russo, Federica, and Paolo Vineis. 2016. ‘Opportunities and Challenges of Molecular Epidemiology’. In Philosophy of Molecular Medicine, edited by Giovanni Boniolo, Marco J. Nathan, London: Routledge. Schulte, Paul A. 1993. ‘A Conceptual and Historical Framework for Molecular Epidemiology’. In Molecular Epidemiology: Principles and Practices, edited by Paul A. Schulte and Frederica P. Perera, San Diego: Academic Press, 3–44. Schulte, Paul A., and Frederica P. Perera, eds. 1993. Molecular Epidemiology: Principles and Practices. San Diego: Academic Press. Shin, Eun Kyong, Ruhi Mahajan, Oguz Akbilgic, and Arash Shaban-Nejad. 2018. ‘Sociomarkers and Biomarkers: Predictive Modeling in Identifying Pediatric Asthma Patients at Risk of Hospital Revisits’. Npj Digital Medicine 1 (1): 50. https://doi.org/10.1038/s41746-018-0056-y. Smith, Martyn T., Kathryn Z. Guyton, Catherine F. Gibbons, Jason M. Fritz, Christopher J. Portier, Ivan Rusyn, David M. DeMarini, et al. 2016. ‘Key Characteristics of Carcinogens as a Basis for Organizing Data on Mechanisms of Carcinogenesis’. Environmental Health Perspectives 124 (6): 713–21. https://doi.org/10.1289/ehp.1509912.

62

F. RUSSO

Strimbu, Kyle, and Jorge A. Tavel. 2010. ‘What Are Biomarkers?’: Current Opinion in HIV and AIDS 5 (6): 463–66. https://doi.org/10.1097/COH. 0b013e32833ed177. Valles, Sean A. 2019. Philosophy of Population Health Science: Philosophy for a New Public Health Era. Routledge. Vineis, Paolo. 2020. ‘Life Trajectories, Biomedical Evidence, and Lessons for Policies’. Frontiers in Public Health 8 (May): 160. https://doi.org/10.3389/ fpubh.2020.00160. Vineis, Paolo, Mauricio Avendano-Pabon, Henrique Barros, Marc ChadeauHyam, Giuseppe Costa, Michaela Dijmarescu, Cyrille Delpierre, et al. 2017. ‘The Biology of Inequalities in Health: The LIFEPATH Project’. Longitudinal and Life Course Studies 8 (4). https://doi.org/10.14301/llcs.v8i 4.448. Vineis, Paolo, and Robert Barouki. 2022. ‘The Exposome as the Science of Social-to-Biological Transitions’. Environment International 165 (July): 107312. https://doi.org/10.1016/j.envint.2022.107312. Vineis, Paolo, and Marc Chadeau-Hyam. 2011. ‘Integrating Biomarkers into Molecular Epidemiological Studies’: Current Opinion in Oncology 23 (1): 100–105. https://doi.org/10.1097/CCO.0b013e3283412de0. Vineis, Paolo, Marc Chadeau-Hyam, Hans Gmuender, John Gulliver, Zdenko Herceg, Jos Kleinjans, Manolis Kogevinas, et al. 2017. ‘The Exposome in Practice: Design of the EXPOsOMICS Project’. International Journal of Hygiene and Environmental Health 220 (2): 142–51. https://doi.org/10. 1016/j.ijheh.2016.08.001. Vineis, Paolo, Cyrille Delpierre, Raphaële Castagné, Giovanni Fiorito, Cathal McCrory, Mika Kivimaki, Silvia Stringhini, Cristian Carmeli, and Michelle Kelly-Irving. 2020. ‘Health Inequalities: Embodied Evidence Across Biological Layers’. Social Science & Medicine 246 (February): 112781. https://doi. org/10.1016/j.socscimed.2019.112781. Vineis, Paolo, Phyllis Illari, and Federica Russo. 2017. ‘Causality in Cancer Research: A Journey through Models in Molecular Epidemiology and Their Philosophical Interpretation’. Emerging Themes in Epidemiology 14 (1): 7. https://doi.org/10.1186/s12982-017-0061-7. Vineis, Paolo, and Michelle Kelly-Irving. 2019. ‘Biography and Biological Capital’. European Journal of Epidemiology 34 (10): 979–82. https://doi.org/10. 1007/s10654-019-00539-w. Vineis, P., and F. Perera. 2007. ‘Molecular Epidemiology and Biomarkers in Etiologic Cancer Research: The New in Light of the Old’. Cancer Epidemiology Biomarkers & Prevention 16 (10): 1954–65. https://doi.org/10.1158/ 1055-9965.EPI-07-0457.

SOCIO-MARKERS AND INFORMATION TRANSMISSION

63

Vineis, Paolo, and Federica Russo. 2018. ‘Epigenetics and the Exposome: Environmental Exposure in Disease Etiology’. In Oxford Research Encyclopedia of Environmental Science, edited by Paolo Vineis and Federica Russo. Oxford University Press. https://doi.org/10.1093/acrefore/978019 9389414.013.325. WHO. 2022. ‘Social Determinants of Health’. In Health Topics. https://www. who.int/health-topics/social-determinants-of-health#tab=tab_1. Wild, Christopher P. 2005. ‘Complementing the Genome with an “Exposome”: The Outstanding Challenge of Environmental Exposure Measurement in Molecular Epidemiology’. Cancer Epidemiology Biomarkers & Prevention 14 (8): 1847–50. https://doi.org/10.1158/1055-9965.EPI-05-0456. ———. 2012. ‘The Exposome: From Concept to Utility’. International Journal of Epidemiology 41 (1): 24–32. https://doi.org/10.1093/ije/dyr236. Yousef, Malik, Mohamed Ketany, Larry Manevitz, Louise C. Showe, and Michael K. Showe. 2009. ‘Classification and Biomarker Identification Using Gene Network Modules and Support Vector Machines’. BMC Bioinformatics 10 (1). https://doi.org/10.1186/1471-2105-10-337.

What’s Wrong with the Biologization of Social Inequalities in Health? A History of Social Epidemiology and Its Moral Economy of Objectivity Mathieu Arminjon

1

Introduction

The existence of social inequalities in health has been in evidence since at least the nineteenth century and was brought to light by the great public health figures of the twentieth century. However, it was not until the 1960s and 1970s that a discipline emerged, “social epidemiology” (Krieger 2001), specifically dedicated to studying “the social distribution and social determinants of states of health” (Berkman and Kawachi

Funding: This research was supported by a grant from the Swiss National Science Foundation (SNSF 200460). M. Arminjon (B) School of Health Sciences (HESAV), University of Applied Sciences and Arts Western Switzerland (HES-SO), Lausanne, Switzerland e-mail: [email protected]

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 É. Giroux et al. (eds.), Integrative Approaches in Environmental Health and Exposome Research, https://doi.org/10.1007/978-3-031-28432-8_4

65

66

M. ARMINJON

2000, 6). There is no doubt that the discipline became more prominent more recently with the advent of the post-genomic era, which fostered the development of epigenetics and exposomics. The “omics” sciences have, in fact, opened the way for non “genocentric” perspectives, which seek to find molecular biomarkers of exposure (Shostak and Moinester 2015). Social epidemiology might be no longer limited to recording statistical correlations between exposure factors and morbidity or mortality rates. We might now be able to “unpack the black box” of the risk-factor epidemiology (Prior et al. 2018). “Omics” sciences might, in fact, complement classical social epidemiology by providing it with a molecular basis (Vineis, Delpierre, et al. 2020). Public health strategies could finally lean on the biomedical knowledge that it has historically lacked. These potential developments have been hotly debated (Niewöhner 2011; Lock 2013; Meloni 2015), and sometimes criticized, by social scientists. In academia, resources are limited and access to them is subject to competition. In the “truth regimes” of the biomedical sciences, biological, or, more precisely, molecular evidence confers a symbolic advantage (Rose 2001; Bliss 2015). In the current “biopolitics of postgenomics” (Bliss 2015, 188), the biologization of health inequalities could quite simply lead to relegating epidemiological evidence that, until now, was sufficient for objectifying social inequalities in health and stimulating public health measures, into the background. Worse, their biomedicalization might contribute to individualizing and thus depoliticizing the social issues addressed by social epidemiologists. One could claim that the biologization of the relationship between health and the social environment constitutes an entry into a “neoliberal regime of truth” (Dupras and Ravitsky 2016) that valorizes biomedical technoscience and individualism over epidemiological methods and a representation of health at the population level. Without minimizing those dangers, we seek here to evaluate to what extent the biologization of research on health inequalities represents a risk of depoliticization. I propose here an epistemological and political history of social epidemiology, especially on a branch of this discipline that is grounded on a psychosocial etiological model. I establish an intellectual filiation between the key actors who played a main role in the development of the discipline. This lineage leads us from the work of Walter B. Cannon on the physiology of stress, to the allostatic load model. The latter is one of the most representative models of contemporary social

WHAT’S WRONG WITH THE BIOLOGIZATION OF SOCIAL …

67

epidemiology aiming to address how socioeconomic factor leaves biological imprints in our bodies by exposure to chronic psychosocial stress (Vineis, Avendano-Pabon, et al. 2020). My aim is to show that psychosocial epidemiology is a specific branch of research on social inequalities in health. Unlike other approaches, such as materialist epidemiology for instance, for which health inequalities are exclusively and directly determined by material deprivation, psychosocial epidemiology is based on a specific pathophysiological model centered on psychosocial stress.1 I then propose to show that in its historical foundations as well as all along its development, psychosocial epidemiology is marked by a “moral economy of objectivity”. Drawing on Didier Fassin (2009), I mean here that despite the different historical and geographical contexts in which psychosocial epidemiology grown, the work of psychosocial epidemiologists is framed by a set of common values that bind together the production of objective data (clinical, epidemiological and biological) and an emotional revolt against social inequalities. Then, to social epidemiologists, biologization provides a gain of objectivity that plainly participates in the struggle against social inequalities in health. This history of the field thus leads me to qualify the novelty as well as the seriousness of the risks associated with the biologization of social inequalities in health. I conclude that the main risk of depoliticization of social inequalities might not rest on objectification methods but on the social factors that tend to make data invisible.

2

Walter Cannon and the Social Etiology of Disease

Walter Cannon is mainly known for his work on strong emotional reactions and for having introduced the notion of homeostasis. His laboratory work has shown that in situations of danger or pain, the body prepares

1 By materialist approach to social inequalities in health, we mean the explanatory models used in the Black Report (Black 1980) as well as the “neo-material interpretation”. According to Lynch, “The neo-material interpretation asserts that health inequalities result from the differential accumulation of exposures and experiences that have their sources in the material world” (Lynch et al. 2000, 1202). For discussions of these approaches see Macintyre (1997), West (1998), and Marmot and Wilkinson (2001). We return to this debate in the last section of the chapter.

68

M. ARMINJON

for “fight and flight” through sympathetic activity (Cannon 1922). These situations trigger a release of adrenaline that promotes defense reactions such as blood clotting, the delivery of sugar into the blood, and an increase of blood flow. But adaptive mechanisms are not only present in dangerous situations. The organism is endowed with a set of compensatory mechanisms allowing it to maintain the stability of its internal environment and, therefore, to maintain a functional normality. In 1926, Cannon coined the word “homeostasis” to describe “the coordinated physiological processes that maintain stable most of the states in the organism” (Cannon 1926, 19) such as the content of water, salt, and sugar of the blood or the constancy of the internal temperature. The organism is thus endowed with regulatory mechanisms allowing it to maintain functional safety margins and thus to stay “normal”. Cannon’s contributions to physiology and medicine, though, should not be reduced to “fight or flight” mechanisms and homeostasis. His work must be viewed as part of his intellectual development as well as in his social and political context. Cannon’s experimental work on emotions and homeostasis is closely related to the question of the emotional etiology of diseases (Cannon et al. 1911; Cannon 1916). Notably, Cannon was not simply content to approach this theme from a strictly physiological perspective. With a medical background, Cannon was careful to use clinical examples drawn from his colleague’s experiences. If some examples he refers to are related to family, or even subjective matters, the socioeconomic theme is striking. For instance, Cannon refers to a case involving a fracture that does not heal as long as the patient shows concern for his family’s situation; a hospitalized sixty-five-year-old diabetic businessman with a controlled diet who presents a sudden increase in blood sugar level when he learns that his employer wants to fire him; a patient whose vomiting only ceases when the doctor discharges a tax debt he owes to the tax collector; etc. (Cannon 2020 [1930]). These pathologies, which he associates with an overactivation of the “sympathico-adrenal” system, raise a scientific problem. How do we establish that these emotional reactions— which constitute a selective advantage from a Darwinian perspective—can both be protective in a situation of danger, yet, under certain conditions, result in diseases associated with the deregulation of homeostatic stability? This question occupies his research from the 1930s onwards. Two papers are significant in this respect. In “Stress and Strain in Homeostasis” (Cannon 1935), Cannon first notes that sympathectomized animals

WHAT’S WRONG WITH THE BIOLOGIZATION OF SOCIAL …

69

are no longer able to compensate for “stress” such as cold, lack of oxygen, lack of blood sugar, or blood loss by physiological regulation (strain) measures. He deduced it would be possible, in principle, to study the strength and endurance of the organism’s self-regulating capacities: “where the limits lie beyond which stresses overwhelm these corrective factors and significantly alter the steady state of the internal environment” (Cannon 1935, 7). Better still, we might be able to measure the “vitality” and thus the risk of disease of an individual by means of a “vital index” (Cannon 1935, 7) taking into account the conditions likely to alter it—“prolonged labour, fatigue, the demands of school”—at different key periods of life such as “childhood, adolescence and old age” (Cannon 1935, 14). In “The Role of Emotion in Disease”, Cannon details epidemiological observations: infections have decreased in the modern world in favor of “nervous strains” (Cannon 1936, 1453). Tuberculosis now kills less than the automobile. About 60% of the population live in cities where specialized work is monotonous and job loss adds to daily anxieties. Suicide rate increased with the Great Depression. Cardiovascular diseases tripled in 30 years. Exophthalmic goiters doubled between 1906 and 1910. Data showed a “shift in the etiology of disease” (Cannon 1936, 1455). Clinicians, Cannon thought, should admit that “the seriousness of infection has been undergoing a remarkable decline, and strains and stress, especially affecting the nervous system, have been on the increase” (Cannon 1936, 1454). If an explanation of the emotional etiology of these diseases progressed—in part thanks to his own research in this area—, Cannon regrets that “reliable statistics are hard to obtain” (Cannon 1936, 1454). In other words, the burden of proof would benefit from a combination of experimental, clinical, and epidemiological data. It is important to note that the word “stress”—understood as the ongoing sense of external, mainly social, events imposing a physiological and emotional strain on the organism—was already in use both in everyday language and in medical literature (Jackson 2013). Also, Cannon, by linking the effects of certain social determinants of health to a set of physiological systems, or even to certain hormones such as adrenaline, is one of the first, if not the first, to biologize and molecularize stress. Keeping in mind the fact that biomedicalization is potentially a by-product of a “neoliberal regime of truth”, we could even go further by comparing this biologization to the liberal position that some associate him with based on the last chapter of his book The Wisdom of the

70

M. ARMINJON

Body (Cannon 1932), entitled “Epilogue: Relations of Biological and Social Homeostasis”. Cross and Albury consider that “he is explicitly opposed to the kind of central planning one finds in socialist or communist economies, but he also retains the classic liberal notion that the economy can function as a self-regulating system, even if not in pure laissez-faire form” (Cross and Albury 1987, 176). However, this reading conflicts with Cannon’s interest in the socio-emotional causes of disease. Similarly, it does not ring true as regards the epilogue of the book, in which Cannon argues that the economic crises that marked the economic activity of the United States until the crash of 1929 are proof of the absence of social homeostasis, and that political authorities should be inspired by the homeostatic mechanisms of the biological body in order to guarantee the stability of the political body. In this sense, he praises the interventionist policies implemented by socialist and communist governments. To him, these interventionist methods prove that “intelligence applied to social instability can lessen the hardships which result from technological advances, unlimited competition and the relatively free play of selfish interests” (Cannon 1932, 303). But what does “intelligence applied to social instability” mean for Cannon? Homeostasis must be re-contextualized as part of the reformatory movement initiated by several intellectuals during the inter-war period, which sought to reform the political organization of the United States and defend a system of social protection (Tournès 2011). On this regard Cannon can be considered as a scientist and an activist (Kuznick 1987; Arminjon 2016). To be fully understood, Cannon’s views on biology and politics must be put in context with those of his contemporaries, especially with those of eugenicists. When Cannon wrote The Wisdom of the Body, he had been in contact with Charles Richet for several years. Richet, a French physician and physiologist, winner of the Nobel Prize in 1915, is known for his work on anaphylaxis, but also for a series of books in which he expresses his concern about the declining vitality of the French, who are supposedly unable to face the challenges of the modern world. The Great War, he said, “made a backward selection”, the “brave, the young, the strong, the vigorous, the beautiful” died, leaving only the “human waste” (Richet 1919, 44, my translation). This ardent defender of rightwing anarchism believes that state intervention should be reduced to a minimum except in the context of reproduction. It must favor the reproduction of the most suitable French people by preserving the “white race”

WHAT’S WRONG WITH THE BIOLOGIZATION OF SOCIAL …

71

from any mixing with the “others”, which he considers inferior. This argument would be considered completely ludicrous if Richet had not relied solely on it to provide his eugenicist and deeply racist solution to this perceived problem. Two radically different political uses of biology are opposed here. Eugenicists seek a type of governance whose intervention is limited to the implementation of biopolitical techniques to “select” organisms capable of facing the challenges of modernity. To Cannon, a physiologist must consider the type of governance that organizes social life in such a way as not to subject organisms to constraints that would take them above or below the homeostatic physiological limits beyond which diseases appear. Here, we come very far from a liberal vision—Cannon is, on the contrary, explicitly opposed to laissez-faire liberalism—in which biologization would serve as an individualistic vision of the body and its pathologies. The identification of physiological limits responds to the need for a mode of governance that intends, upstream, to prevent the onset of economic crises and diseases by intervening in their structural causes, and, downstream, to cushion their physiological cost, mainly by unemployment insurance and health insurance. We shall see that it is the criticism of homeostatic theory—paradoxically based on partial readings of Cannon’s theses—that proved conducive to the development of the psychosocial approach to social inequalities in health. But if the question of the connection between Cannon and his successors can be debated on an epistemological level, we can observe continuity in the moral and political values structuring their scientific practices.

3 René Dubos: Stress and Susceptibility to Infectious Diseases In 1959, René Dubos, an agronomist, biologist and ecologist of French origin, who made his career in the United States, published The Mirage of Health (Dubos 1959). In large part, he defended positions closely aligned with the “McKeown thesis” (McKeown and Brown 1955), according to which the demographic explosion, the decrease in infant mortality, and, ultimately, the increase in life expectancy that began at the turn of the twentieth century are not due to the success of scientific and clinical medicine. Both Thomas McKeown and Dubos relied on epidemiological data indicating that tuberculosis mortality—considered as the main

72

M. ARMINJON

cause of death at the turn of the twentieth century—declined well before the discovery of the bacillus in 1882 and well before the generalization of treatments in the post-war years. But contrary to McKeown, who only attributed this change to the rise in living standards (mainly access to more and better food), Dubos claimed that the sanitary measures put in place by hygienists significantly contributed toward eradicating the socioenvironmental causes of diseases that emerged during the industrial era. Dubos highlighted a paradox: we incorrectly attribute the results of social and preventive medicine to scientific medicine, especially to Louis Pasteur and Robert Koch’s bacteriology. Although social and preventive medicine may be perceived as responsible for the general improvement of health, its successes were not based on any scientific foundation, but on a naïve “crusade for pure air, pure water, pure food” (Dubos 1965, 164). By identifying this accidental effectiveness, Dubos intended to reveal what he considered to be the erroneous assumptions of scientific medicine. At the very moment when Koch gave his famous lecture reporting the discovery of the bacillus, Dubos notes, several people in the assembly were probably healthy carriers (Dubos 1959, 104). The whole population was not equal in front of tuberculosis since mortality dropped at first in the wealthiest classes. The living conditions of the working class— poverty, crowding, poor housing, lack of access to quality and abundant food, working conditions, exposure to pollutants, etc.—made the bodies of the poorest people more susceptible to tuberculosis. The idea of social inequality in the face of the risk of developing disease is indissociable from the critique of the etiological model—monocausal and specific—which, according to Dubos, is characteristic of modern medicine. If Pasteur’s and Koch’s approaches was scientifically based—while the hygienists’ were ideological—it was nevertheless founded on an incorrect view of infectious diseases, which considers the external pathogen to be the only cause of disease, i.e., one specific germ induces one specific disease: The sciences concerned with microbial diseases have developed almost exclusively from the study of acute or semi-acute infections caused by virulent microorganisms acquired through exposure to an exogenous source of infection. In contrast, the microbial diseases most common in our communities today arise from the activities of micro-organisms that are ubiquitous in the environment, persist in the body without causing obvious harm under ordinary circumstances, and exert pathological effects only when the infected person is under conditions of physiological stress. (Dubos 1965, 164)

WHAT’S WRONG WITH THE BIOLOGIZATION OF SOCIAL …

73

In the 1950s, Dubos came to the “agent-host-environment” model for which the susceptibility of the host depends on their physiological state modulated by the “physiological stress” they are exposed to, which, following Cannon, is understood in both a physical and psychological sense (Dubos 1952). For Dubos, there is no doubt that the workers’ material and social conditions in the context of the Industrial Revolution made these populations more vulnerable to tuberculosis than the wealthier classes. Dubos’ research above all laid the foundations for his aspirations for future medicine, which could incorporate findings of the past and reconcile the environmental and humanist vision of the hygienists with virologists’ experimental approach (Dubos 1959, 218). Following the example of hygienism, which was concerned with revealing the social factors specific to industrial societies, “environmental medicine” (Dubos 1963, 6) should seek to uncover the main sources of stress specific to modern societies. But unlike hygienism, this environmental medicine must be scientifically based by systematically studying the “correlations between the disease processes that will occur in years to come in given human groups, and the various forms of stress to which these groups have been subjected in the past” (Dubos 1963, 7). This epidemiological project is evidently reminiscent of Cannon’s, but it implies important shifts with respect to the homeostatic model. By introducing the agent-host-environment model, Dubos sought to move away from anhistorical representations of diseases—both the one that accompanies the “cause specific” model, for which a single agent constitutes the cause of a specific disease, and the homeostatic model of pathologies that posits normality to be a fixed state. If pathologies depend on the susceptibility of the host at a given moment, itself determined by a particular socioeconomic context, then we must conclude that “each civilization has its own characteristic pattern of diseases, determined by its social structure and its way of life” (Dubos 1963, 1). Such a conception calls for readjusting the homeostatic model to take account for time, be it the history of the individual or the history of a civilization. Dubos notes that humans show a capacity for biological and technological adaptation to the microorganisms present in the environment, but also to the other “stresses” of modern life such as pollutants or overpopulation. Yet, adaptation does not mean that the primitive defense mechanisms of the

74

M. ARMINJON

organism are no longer active in modern societies. Rather, their inappropriate activation for these adaptation purposes potentially presents a long-term danger: The properties of civilized life now deter man from either fighting or running away. It is not unreasonable to assume that the misguided and misspent reaction mobilized by the nervous and humoral mechanisms of the fight and flight response leave on the body and the mind scars which become apparent only as they accumulate. (Dubos 1963, 5, my emphasis)

The reference to Cannon’s work is obvious. However, Dubos intends to introduce two new notions that are absent in Cannon’s work. Firstly, the chronic activation of the sympathico-adrenal system for adaptation purposes has a cost, it leaves homeostatic “scars”. If these are harmless at the beginning of life, their cumulative effects can impact health several years later. But the critique of a fixed homeostatic stability also leads Dubos to insist on adaptive plasticity to an environment especially in childhood during which inherited homeostatic responses are modulated by the environment. The thesis of a shaping of homeostatic responses by the environment—which he qualifies as “biological Freudianism” (Dubos 1974, 98; Dubos et al. 2005 [1966])—, along with homeostatic scars, calls for the abandonment of a fixist and passive representation of homeostasis in order to emphasize an overall “biological creativity” (Dubos 1974, 97). In many respects, Dubos does not exactly break with Cannon, he in fact highlights some aspects of the homeostatic model, which he places in a historical context. For, if Cannon had in mind the idea that health is determined by social factors, Dubos insists on adaptation over time: if resistance and vulnerability to disease depend on social conditions, then physiology and pathology are no longer the exclusive domain of the natural sciences, but of the historical-social sciences too. And if the conditions that favor or not the appearance of specific pathologies depend on socio-historical contingencies, then they are preventable. This precise point brings us back again to the status of the activist-scientist. In this respect, Dubos is explicit: It is not impossible that in the future, as in the past, effective steps in the prevention of disease will be motivated by an emotional revolt against some

WHAT’S WRONG WITH THE BIOLOGIZATION OF SOCIAL …

75

of the inadequacies of the modern world (…) This attitude need not mean a retreat from science – far from it. (Dubos 1959, 219, my emphasis)

Indeed, it is precisely the fact that both physiology and epidemiology demonstrate that some diseases are preventable by means of social measures, which leads to an internal feeling of revolt among some scientists. And, it is with reference to Dubos, that the first generation of social epidemiologists came to embody this revolt and bring about the accompanying scientific program.

4 John Cassel: Incarnating the Emotional Revolt Against Diseases Among those considered as pioneers in social epidemiology, John Cassel occupies a special position. As Leonard Syme puts it: “He was doing some of the very best work at the time and, interestingly, much of his research is still the best. Without doubt, his persistent influence is also due to his theoretical contribution” (Syme 2005, 5). And the direct intellectual link between Dubos and Cassel is evidenced in Cassel’s final paper, “The Contribution of the Social Environment to Host Resistance” (Cassel 1976). Cassel’s input overlaps with epidemiological research but also theoretical and methodological thoughts that appear to be crucial for understanding how the host-agent-environment model offers a renewed, psychosocial representation of etiology. Cassel notes that the concept of stress is ambiguous. For instance, authors such as Hans Selye contributed to the understanding of mechanisms of stress, but have also made it more complex, notably by qualifying stress as a physiological state (where Cannon spoke of strain) and the environmental conditions as stressors (where Cannon spoke of stress). But most important for Cassel is the recognition that most stress theories promote the idea that “a stressor [would be] capable of having a direct pathogenic effect analogous to that of a physicochemical or microbiologic environmental disease agent” (Cassel 1976, 109). Therefore, just as it has been assumed that a specific agent induces a specific disease, it would also be wrong to think that a specific socioenvironmental situation causes a specific disease. In line with Dubos, Cassel claims that environmental situations act “indirectly” as a “conditional stressor”, which, “by altering the endocrine balance in the body, increase the susceptibility of the organism to direct noxious

76

M. ARMINJON

stimuli, i.e. disease agents” (Cassel 1976, 109). Inspired by Harold G. Wolf, Cassel distinguishes psychosocial stress from an external agent of a physico-chemical nature which directly impacts the organ structure, physically and chemically. Socioenvironmental stresses are not the direct cause of disease. They act as “signal” or “symbol” (Cassel 1974, 473; 1976, 111) modifying the endocrine balance, thus leading to a susceptibility to disease. This distinction is central for understanding to what extent the “psychosocial model” departs from the traditional medical approach of etiology. Using the term “psychosocial” here may seem paradoxical since it refers apparently to the individual and not the social determinant. We have little difficulty imagining the same emotional symbol impacting the life of one individual, while leaving another indifferent. Cassel sees this as a problem of abstraction that is of little importance to the epidemiologist, for whom it is important to observe whether “the same relationships or social circumstances within a given culture (or, perhaps, subculture) regularly produce such a class of signals” (Cassel 1976, 111). In other words, there must be mechanisms that induce, in one group and not in another, a state of susceptibility to disease. To Cassel, the issue of disease susceptibility calls for a departure from the usual biomedical strategy of looking for the external agent or socioenvironmental situation that induces a specific pathology. It would be a more logical approach, to examine all disease outcomes related to exposure to the postulated stressor or stressors or alternatively to identify subsets of the population who by virtue of their personal or environmental characteristics are known to be at high risk of specific clinical manifestations and examine the role of psychosocial stressors in facilitating the appearance of those manifestations. (Cassel 1974, 474)

Cassel helped clarify Dubos’ theses and establish the methodological foundations of “environmental medicine”. He also undertook experimental research seeking to operationalize it, i.e., by identifying the psychosocial factors that promote susceptibility to disease. Cassel conducted research on the effects on health of the rapid social and cultural changes brought about by industrialization. For example, he showed that in a North Carolina population the first generation of workers who went from being farmers to factory workers was in poorer

WHAT’S WRONG WITH THE BIOLOGIZATION OF SOCIAL …

77

health than the second generation who only knew the factory (Cassel and Tyroler 1961; Cassel et al. 1960). In other words, cultural and social transitions increase rates of psychological and somatic ill health. Some commentators limit Cassel’s contribution to the study of the effect of large, abrupt social changes on health (Krieger 2001). However, Cassel played a central role in the study of other social factors, such as social disorganization and its effects on health, particularly on the AfricanAmerican population in the United States. This point is even more critical as it demonstrates how Cassel not only operationalizes Dubos’ scientific program, but also exemplifies an “emotional revolt” at the source of knowledge production on social inequalities in health. Indeed, the information available on Cassel’s background—though limited—confirms his personal concern with racial discrimination. Like other historical figures of the field,2 Cassel was a native of South Africa, where he defended the development of community clinics for the Zulu population, which was kept on the margins of South African society. The onset of apartheid in 1948 forced him and his wife into exile in 1953 (Geiger 2002). His research activity, especially as a director of the Department of Epidemiology at the School of Public Health of the University of North Carolina, is part of a life of commitment and “emotional revolt”. This was born out of his contact with the Zulu population in South Africa and continued in the United States during the 1960s and 1970s in the context of the civil rights movements. This commitment is recurrent in Cassel’s publications, where the question of blood pressure is seen as a central problem both from a political and an epistemological point of view. The question of blood pressure in the African-American community is the source of an “emotional revolt”, but it also represents the kind of medical issue psychosocial epidemiology must tackle. It has been known since the 1920s that the members of the AfricanAmerican community suffer from a particularly high rate of morbidity and mortality (Arminjon 2020a). Although the causes of these social inequalities in health have long been debated, studies have shown that these differences can hardly be explained by genetic factors or, paradoxically, by the poverty affecting the community alone. Cassel and his collaborators studied stroke mortality, specifically in the black community, using 2 Key figures in public health and social epidemiology were banished from South Africa for the same reasons (Arminjon 2020a). Zena Stein and Mervin Susser were forced into exile in 1955, Emily and Sidney Kark in 1958.

78

M. ARMINJON

data from the North Carolina State Department of Health (Neser et al. 1971). They showed that the stroke rate, which is particularly high in the black population, did not correlate with poverty. They also revealed that mortality rates were greatly associated with the level of social disorganization measured by the Harvey Smith Index. This included the percentage of single-parent families, illegitimate births, the rate of men sentenced to prison camps, the percentage of the population separated or divorced, and the percentage of children under the age of 18 not living with both parents. In the white community, the level of social disorganization did not correlate with the incidence of cardiovascular disease. Cassel concludes that: Perhaps these findings reflect the subservient role that Blacks (until perhaps recently) have been forced to occupy in our society. The second speculation is the possibility that in the face of social disorganization Whites have more resources, including sources of social support to help buffer their physiologic processes from these effects. (Cassel 1976, 116)

Although there is no doubt that the African-American population is economically dominated, material deprivation appears not to be the only determining factor in susceptibility to cardiovascular disease. Here we see the psychosocial indirect explanatory model at work: the social status of economically dominated people places them in contexts of social disorganization that prove to be both a source of stress in themselves and factors that diminish their other resources, such as social support, that may be protective against stress. At the interface between the theory of disease susceptibility induced by psychosocial stress and the issue of racial inequalities in health, social epidemiology would follow two main lines of research. One, through Leonard Syme and Michael Marmot, concerns the “deracialization” of biostatistics related to social inequalities in health—this aspect played a crucial role in the internationalization of social epidemiology. The other involved the sophistication of the pathophysiological model which, through the work of Peter Sterling and Joseph Eyer, led to the development of the concept of allostasis. It is at the intersection of these two lines of research that the allostatic load model emerged.

WHAT’S WRONG WITH THE BIOLOGIZATION OF SOCIAL …

79

5 From Syme to Marmot: The Britannicization and the Internationalization of Social Epidemiology Leonard Syme occupies a key place in the development of social epidemiology. He conducted his first research in the early 1960s, taking advantage of the vast network of research on coronary heart disease (CHD), mainly represented by the Framingham study. He analyzed data (Syme et al. 1965) from a large study of the geographic distribution of CHD mortality in the United States: very high rates on the East and West coasts and in the Detroit-Chicago metropolitan area, and low rates elsewhere. In line with Cassel’s work, he demonstrated that, given equal risk factors (smoking, blood pressure), rates of CHD were higher among men who have changed jobs, moved from one geographic area to another, particularly concerning those who have moved from agricultural jobs to urban jobs. Syme explicitly embraces the conceptual psychosocial model inherited from Dubos via Cassel, and the topics studied by Cassel, such as the health effects of sudden socioeconomic transformations. He also considers the social causes that might help understand the high rate of cardiovascular disease affecting the African-American community. On this last point, Syme et al. published “Social Class and Racial Differences in Blood Pressure” (Syme et al. 1974) in which they presented the results of research seeking to clarify the influence of environment on hypertension. Involving 22,078 people, both white and black, the data confirms that hypertension is always higher in the black community than in the white community. But when the distribution is analyzed by social class, it appears that the most disadvantaged in the black population, as well as the poorest in the white community, albeit in a lesser proportion, always have higher blood pressure than members of the most advantaged social classes, regardless of their ethno-racial community. The authors thus reveal an inverse correlation between blood pressure level and socioeconomic status: The social class gradient in blood pressure among blacks suggests that an environmental influence may be involved in determining blood pressure; the fact that an identical social class gradient was observed among whites suggests that something related to lower social class position may be involved beyond any racial influence. (Syme et al. 1974, 619)

80

M. ARMINJON

This quote is insightful for it refers to the notion of “social class”, whereas the studies mentioned so far have compared similar populations in contexts of major and sudden socioeconomic transformation or have mobilized racial categories in a preferential manner. Perhaps Syme’s initial training in sociology, not medicine or epidemiology, has something to do with this shift in focus from race to class. In any case, it is important to note that this effort to subsume racial classes within social classes, i.e., to make the former special cases of the latter, helps broaden the visibility of disparities and leads to him introducing the idea of a determination of health that applies “beyond any racial influence”. In other words, Syme used the notion of “gradient” that will be more widely popularized by Michael Marmot under the term “social gradient in health”. In fact, the two researchers worked with each other. Marmot, a medical graduate, completed a doctorate at the Berkeley School of Public Health under the direction of Syme. His thesis focused on the risk of coronary heart disease among Japanese people living in California, which was five times higher than those living in Japan (Marmot and Syme 1976). He showed that Japanese people living in California who had maintained a traditional lifestyle did not have a higher risk of coronary heart disease. But above all, Marmot played a key role in promoting the psychosocial model in the Britannic public health culture, where research on social inequalities had mainly employed a materialist approach. Marmot finished his doctorate in 1976 and joined the London School of Hygiene and Tropical Medicine, where Geoffrey Rose conducted the Whitehall study from 1967 to 1969, which aimed to analyze the prevalence of cardiovascular disease and mortality in a cohort of more than 1800 British male civil servants aged 40–64 years (Marmot et al. 1978). Whitehall I3 was not designed to investigate differences in health by social class but rather individual lifestyle factors, such as early myocardial ischemia, which is a predictor of CHD mortality. Marmot classified the

3 From 1985 to 1988, Marmot conducted the Whitehall II Study to better investigated the psychosocial causes of the social gradient in health. The second cohort study included 10,308 civil servants aged 35–55 from both sexes.

WHAT’S WRONG WITH THE BIOLOGIZATION OF SOCIAL …

81

data according to the 4 Registrar Occupational Categories (Other, clerical, professional executive, administrative) and established the existence of a socioeconomic gradient of CHD mortality4 : In the Whitehall study, 17,530 civil servants were classified according to employment grade, and their mortality was recorded over 10 years. There was a steep inverse relation between grade and mortality. (Marmot et al. 1984, 1003)

In doing so, Marmot not only identified the existence of a socioeconomic gradient but also imported the etiological model of psychosocial stress and host susceptibility from the US public health culture, dominated by racial concerns, into the British one. To view this in context, we need to turn our attention to the state of research on health inequalities in the United Kingdom in the mid-1970s. Research on health inequalities in the United Kingdom dates back at least to the work of Edwin Chadwick and Friedrich Engels in the late nineteenth century (Oppenheimer et al. 2002). But it benefited especially from the development of social medicine in the post-war years led by the New Public Health figures such as John Ryle and Thomas McKeown, who inspired another generation, for example Geoffrey Rose and Jerry Morris. But Dorothy Porter has documented how, during the 1950s and 1960s, research on the social determinants was gradually relegated to the background in favor of studies regarding health behaviors (Porter 2011). A new generation of researchers would go on to revive the issue. In 1976, Richard Wilkinson, a young left-wing sociologist at the time, sent an open letter to the Labour Minister in the government, reporting on his analyzes which revealed that, despite health insurance and the general increase in life expectancy for the entire population, the mortality gap between the poorest and richest social categories had increased (Wilkinson 1976). The Minister commissioned a research group, which delivered similar conclusions in the famous Black Report (Black 1980). The authors then adopted a principally materialist explanation, ultimately

4 Marmot notes that it was Syme, to whom he showed the data, who drew his attention to the existence of a socioeconomic gradient, that is, the fact that social inequalities are not only measurable between the most disadvantaged and the most affluent groups, but, more generally, between all grades of the social hierarchy (Marmot 2001, 1165).

82

M. ARMINJON

a variant of the McKeown thesis, which centered on structurally determined health behaviors according to social classes (Macintyre 1997; Marmot 2001). The notion of stress and therefore the psychosocial model, appears only in passing (a reference to the work of Joseph Eyer) and is mostly contested. In this context, Marmot’s work clearly contributed to the importation of the psychosocial model into the UK public health culture. In 1984, he wrote: Social class differences in mortality across a wide range of diseases persist despite general improvements in mortality. The Black report suggested that specific socio-economic features, such as smoking or accidents at work, might explain some of these differences; but other more general social influences must also be operating. Others too have suggested that nonspecific susceptibility, rather than a clustering of specific causes, may be linked with social class. (Marmot et al. 1984, 1003)

The “others” to whom he refers are precisely those who, like Cassel and Syme, embraced the psychosocial model. It is precisely in order to test the psychosocial stress hypothesis that Marmot set up the Whitehall studies II. The study confirmed the existence of a social gradient in health. Against the materialist thesis, which posits that socioeconomic conditions directly determine health inequalities, Marmot and his teams showed again that the poorest die more than the richest, but also that any individual is in poorer health relative to those who occupy a higher place in the hierarchy. But still, beyond the risk behaviors linked the socioeconomic level, such as the higher consumption of alcohol or tobacco in the popular social classes, other factors of a psychosocial nature were confirmed. Mortality, somatic and mental morbidity increases as the level of well-being felt decreases, as workers have less control at work, when their work involves fewer skills, and as they benefit from less social support protecting them from stress, anxiety and depression. Marmot regularly refers to the physiology of stress and, sometimes, to the concept of allostasis (Steptoe and Marmot 2002; Marmot 2003) to support the psychosocial approach, particularly against the materialist approaches (Marmot 2001; Marmot and Wilkinson 2001). Marmot continued his research, notably with Richard Wilkinson, and tended to show that the magnitude of the socioeconomic gradient is correlated not so much to the level of wealth of nations, but to the amplitude of socioeconomic inequalities between the wealthiest and the

WHAT’S WRONG WITH THE BIOLOGIZATION OF SOCIAL …

83

most disadvantaged classes. In the 1990s, he participated in the vast study that the Tony Blair British Labour government commissioned in 1997 and entrusted to Donald Acheson (Acheson and Great Britain 1998). In 2005, Marmot was appointed Chair of the WHO Commission on Social Determinants of Health. In 2008, the commission published a report with recommendations for closing the gap in social inequalities in health within a generation (CSDH 2008). The report showed that the socioeconomic gradient, which Marmot had summarized elsewhere as “the higher the social position, the better the health” (Marmot 2004), should essentially be attributed to the psychosocial stress induced by social inequalities. On this basis, it concluded: “Social injustice is killing people on a grand scale” (CSDH 2008, VII). We have established how some actors in social epidemiology ensured that the psychosocial model was at the forefront of international public health strategies. We will now step further back in time in order to understand how the allostatic model allowed for psychosocial stress to be conceptualized as an embodiment of social inequalities in health.

6 Peter Sterling and Joseph Eyer: From Allostasis to Allostatic Overload When Peter Sterling and Joseph Eyer introduced the concept of allostasis in 1988, they intended to lay the foundations of a new physiological paradigm capable of better accounting for the emergence of pathologies (Sterling and Eyer 2005 [1988]). While there is no doubt that this conceptual effort testifies to the vitality of research on the social determinants of health, this important step must be re-contextualized to be viewed as part of the genealogy that we have traced so far. The allostatic paradigm does not exactly break with previous models, especially the homeostatic one, as it proposes a re-articulation of research conducted at the interface of, on the one hand, social and political concerns, and, on the other, neurophysiology. The first salient feature of the allostatic model is the understanding of the effect of stress in relation to specific sociocultural determinants, especially those of capitalist societies, from a critical and comparative perspective (Arminjon 2016, 2020b). Eyer and Sterling (1977, 5) refer to a study that supports the thesis of the sociocultural variability of hypertension. West African hunter-gatherers have low and stable blood pressure throughout their lives (Epstein and Eckoff 1967). However, in pastoral

84

M. ARMINJON

societies—social organizations characterized by a sedentary lifestyle and agricultural economy—the pattern of hypertension is similar, although much less marked, to that of capitalist societies in which blood pressure increases significantly in adulthood and throughout life. As descendants of slaves in the United States share a genetic heritage with West African hunter-gatherers, the absence of hypertension in these latter populations supports the claim that hypertension is not essential (i.e., not a consequence of aging, genetic or other unexplained “natural” factors), neither is African-Americans’ hypertension, the highest in the world, produced by a “racial defect”. For Sterling and Eyer, therefore, cross-cultural epidemiology indicates psychosocial stress specific to industrialized societies is the primary cause of hypertension in the general population and that social and racial stigmatization amplifies this trend in the African-American community. The cross-cultural variation of vital constants, in particular hypertension, is central to distinguishing allostasis from homeostasis. In the homeostatic model, pathology is thought of as a disruption of regulatory mechanisms, in terms of either a hypo- or hyper-function of normal functions around a normal and fixed value. The allostatic model instead takes addiction as a model of the pathological process. Whether it is adaptation to or anticipation of a repeated stimulus, Sterling and Eyer tell us, “When demand and thus arousal become chronic, the brain-body system adapts at all levels of organization…the body becomes dependent on its own catabolic hormones” (Sterling and Eyer 2005 [1988], 641). Hypertension and diabetes, for example, are examples of physiological adaptations: thickening of blood vessel tissue in the case of hypertension or decreased receptor sensitivity to chronic demands in the case of diabetes. This leads to a paradoxical approach to pathology that Sterling explicitly acknowledges: “The allostasis model clearly presents a paradox: people die, but their internal regulatory mechanisms are intact” (Sterling 2004, 51). There is pathology without dysfunction, i.e., exceeding normal values, beyond homeostatic safety margins, is not a condition of pathology, but of adaptability. The process leading to type 2 diabetes or hypertension is therefore quite functional on a physiological level, or even normal, on a biostatistical level. This is what cross-cultural epidemiology reveals: being hypertensive, in the Western population, and a fortiori, in the African-American community, is functionally and statistically normal. If we cannot consider hypertension to be an “inappropriate” physiological

WHAT’S WRONG WITH THE BIOLOGIZATION OF SOCIAL …

85

state, we can nevertheless deem certain social environments to be “inappropriate” (Sterling 2004, 645). If so, the dysfunctional nature of a social environment cannot be judged based on internal organic parameters, which are considered normal in the statistical sense. Only the mortality rate makes it possible, at a population level, to judge this. The allostatic model leads us to two conclusions: the need to rethink both the relationship between the biological and the socio-historical environment and the need to recharacterize the relationship between the normal and the pathological accordingly. First, the cross-cultural epidemiological data employed by Sterling and Eyer testifies to a desire to place physiology in broad macroscopic contexts. Western humans are not naturally hypertensive; their physiology is the product—as well as the fuel—of rich, industrialized, and capitalist societies based on individual competition. Peter Sterling and Joseph Eyer’s biographies testify that the allostatic model emerged in the context of the 1970s, marked by the civil rights struggle and the development of the radical sciences (Moore 2013). This movement was advocated by scientists promoting the development of a science in service of social justice, requiring a reconsideration of the conceptual bases of biology, for example by founding a Marxist biology (Arminjon 2016). Refusing the opposition between the natural and the social of the “bourgeois sciences”, it seeks to recontextualize the materiality of the body in its historical and political environment. Allostasis—as a concept allowing for understanding of how the physiological parameters are shaped by social context—paved the way for Nancy Krieger’s introduction into social epidemiology of the term “embodiment”: “A concept referring to how we literally incorporate, biologically, the material and social world in which we live” (Krieger 2001, 672). The dual explanatory strategy of social epidemiology is subsumed under this single term, i.e., the effort to explore and articulate social pathways (“societal arrangements of power and property and contingent patterns of production”) with biological pathways (“constraints and possibilities of our biologies”) (Krieger 2001, 672).5

5 The influence of anthropology and philosophy in the development of the concept of embodiment is obvious (see for instance Louvel and Soulier 2022). However, Krieger is part of a broader feminist movement in radical science in which the concept of embodiment, or even of allostasis, is precisely used to think about the social shaping of the female or racialized biological body (Haraway 1991; Fausto-Sterling 2004, 2005).

86

M. ARMINJON

Second, the introduction of the concept of allostatic overload can be seen as an exploration of biological pathways as well as an attempt to overcome the paradox of the allostatic model in relation to the normal and the pathological. The homeostatic model, which defines the onset of disease as a quantitative deviation from normal physiological thresholds, assumes a fixity of vital parameters that are incompatible with the thesis of physiological adaptability. Yet, the allostatic adaptability to sociocultural environments makes it difficult to understand the criteria for entering pathology. Bruce McEwen and his collaborators view the concept of “allostasis” in line with the psychosocial tradition as: “the cost of chronic exposure to fluctuating or heightened neural or neuroendocrine response resulting from repeated or chronic environmental challenge that an individual reacts to as being particularly stressful” (McEwen and Stellar 1993, 2093). But they suggest introducing the notion of “allostatic load” to better conceptualize the pathophysiological “cost” of adaptation and the criteria for distinguishing an abnormal from a normal adaptation. However, this concept did not resolve the issue, since allostatic load refers both to the adaptation process and to the increased risk of disease induced by chronic stress. McEwen and his teams then suggested solving the problem by distinguishing two types of adaptation, associated with two types of chronicity, one could be said to be “natural” while the other is “social” (McEwen and Wingfield 2003). In the animal kingdom, for example, hyperphagia before the hibernation period or the cessation of a female’s menstrual cycles in a situation of famine incompatible with conception, involves deep and chronic physiological fluctuations linked to common adaptive capacities. In those cases, chronicity can lead to “type 1 allostatic overload”, i.e., “when energy demand exceeds supply, resulting in activation of the emergency stage of the life history” (McEwen and Wingfield 2003, 2). Type 1 allostatic overload, for instance overeating to prepare the body for a period of starvation, is qualitatively different from situations where the ingestion of food, for example, is not intended to prepare the organism but to compensate, via the reward circuits, for stress caused by modern life. Type 2 allostatic overload “begins when there is sufficient or excessive energy consumption accompanied by social conflict and other types of social dysfunction” (McEwen and Wingfield 2003, 2). The neurophysiological processes (HPA axis, autonomic nervous system) activated during chronic stress, and the adaptive behaviors (overconsumption of alcohol, tobacco, junk food) which, coupled with a sedentary lifestyle, lead to a pre-diabetic

WHAT’S WRONG WITH THE BIOLOGIZATION OF SOCIAL …

87

state, accumulation of fat, atherosclerotic plaques, left ventricular hypertrophy, etc. The latter conditions are no longer a result of environmental adaptation, but rather of a dysregulation of allostatic mediators. The latter can be said to be pathological as it is a physiological state. It is measurable, in principle, through the neuroendocrine markers, which refer to risk or susceptibility to disease. In spirit, the idea is reminiscent of Cannon’s theoretical evocation of a “vital index”. It is precisely the implementation of this type of index that McEwen and his collaborators tested using data from the MacArthur aging studies cohort from 1988 to 1991. The “allostatic load score” that they created groups together ten physiological parameters related to the “wear and tear” caused by cumulative psychosocial stress associated with socioeconomic status. Results show that this score correlates, over time, with a greater decline in cognitive and physical functioning, as well as an increased risk of cardiovascular disease. Research of this type is still being conducted as part of the LifePath research program (Vineis, AvendanoPabon, et al. 2020). Using data from the 1958 British birth cohort, the measurement of 14 biomarkers in blood showed that a higher allostatic load at age 44 was a significant predictor of mortality 11 years later. The authors concluded: “Our findings add some evidence of a biological embodiment in response to stress which ultimately affects mortality” (Castagné et al. 2018, 441).

7 Discussion: Biologizing Social Inequalities in Health as a Specific Moral Economy of Objectivity To some, the entry into the “post-genomic” era is an example of the continuation of “biomedicalization”. It results in biologizing research on social inequalities in health at the expense of epidemiological approaches. By individualizing and depoliticizing health issues, this movement might be an example of a “neoliberal regime of truth” (Dupras and Ravitsky 2016). I propose here to analyze to what extent the actors in the history of social epidemiology share a moral economy of objectivity i.e., common emotional and scientific values that structure their research and struggle against social inequalities in health. This might be a prerequisite before

88

M. ARMINJON

I would reconsider the risk of the depoliticization associated to the biologization of social inequalities in health. In a paper aimed at clarifying the uses of the concept of “moral economy”, Didier Fassin (2009) distinguishes three conceptual evolutions. The concept was coined by the historian Edward P. Thompson (1966), who showed that the peasant revolts of the eighteenth century or the workers’ revolts of the nineteenth century couldn’t be explained by famine alone. Peasant revolts were an emotional reaction against a new caste that, aligning itself with the principles of the liberal “economy of the free market”, no longer respected the “older moral economy” that guaranteed prices that would ensure their subsistence (Thompson 1966, 67). The second contribution is made by James C. Scott, who sought to understand “how the central economic and political transformations of the colonial era served to systematically violate the peasantry’s vision of social equity” (Scott 2006 [1976], 4). Moral economy, here, designates the value system underlying the southeast peasant revolts aimed at enforcing their right to subsistence. Built on principles of justice, redistribution, and dignity, the moral economy of domination determines the expectations of the dominated vis-à-vis the dominant. In both cases, the moral economy is “a tool specifically constituted to think the relations of difference (in time) and inequality” (Fassin 2009, 1250). By introducing it into the field of the history of science, Lorraine Daston (1995) gives the idea a radically different meaning. It refers to the values and affects that legitimize, at a given moment in history, the individual and/or collective qualities that make a scientist legitimate, frame the norms of a scientific sociability, and guarantee the truth value of scientific knowledge. For instance, the “moral economy of quantification” values the precision of the demonstration: the impersonality and impartiality of numbers prevail over theories, even over the adequacy of nature. The “empiricist moral economy” values the credibility of those who witness phenomena, more than their reproducibility. The “moral economy of objectivity” is finally articulated as the values of solidarity and sharing; knowledge is a collective matter, and the individual must step aside and let the recognition of the facts be the prevailing factor. As pointed out by Fassin, by freeing the concept from its initial context, Daston opens the path for a general theory of moral economies but empties it of any reference to social criticism. The history we have traced here gives us the opportunity to reconcile the use of moral economies in the history of science with its

WHAT’S WRONG WITH THE BIOLOGIZATION OF SOCIAL …

89

political relevance. It appears that beyond the scientific and even political divergences, psychosocial epidemiogists share a set of emotions and political values. Indeed, it is not insignificant that the history of social epidemiology is rooted in Cannon’s career, that of a physiologist and activist who saw in social medicine one of the most striking opportunities for “responsible men” to be able to control human ills such as infectious diseases (Cannon 1932, 303). It continues with Dubos, who insists on linking the scientific and political future of the discipline with an emotional “revolt” against the “inadequacies of the modern world”, an issue that was also very much in contention in the social epidemiology of the 1960s and 1970s. Cassel was personally affected by the struggle for civil rights and racial equality in health. The same can be said for Sterling, Eyer, and Krieger who explicitly qualified their scientific activities as scientific activism forming part of a radical or Marxist scientific agenda. Social epidemiology mainly comprises researchers who see themselves as concerned activists-scientists and intend to contribute to struggle against health inequalities in their own way, i.e., with their biomedical expertise. In other words, the history of social epidemiology must be re-contextualized in the history of activism as the struggle for an alternative, non-neoliberal biopolitics. This story is led by scientists for whom the accumulation of clinical, epidemiological, and biological data objectifying social inequalities in health is crucial for the defense of social health justice. It also appears that actors’ shared emotions and political values shape their sense of objectivity and their used of biological proofs. Cannon did not develop the first biological model of stress within the confined limits of his favored discipline, physiology. The aim was to provide an explanation for a set of clinical manifestations along several axes: physiological, epidemiological, and even anthropological. This biologization is therefore not reductionist in the sense that it would result in one form of explanation—biological—prevailing over others. Moreover, the history presented here shows that biology has a very special explanatory status in social epidemiology which does not tend toward de-socializing health problems, on the contrary. Cassel notes that the psychosocial model differs from current medical conceptions according to which physico-chemical pathogens have a direct and harmful effect on the structure and function of organs. In contrast, the psychosocial model sees host susceptibility as mediated by stressful social-environmental conditions (Cassel 1974, 1976). How could the direct exposure model explain the correlation

90

M. ARMINJON

between social processes—such as abrupt social changes, lack of social support, racial stigmatization, etc.—and increased morbidity? Proponents of the psychosocial model hypothesize here that the mediation of a neurophysiological model of stress is required in order to account for this. Making the distinction between direct and indirect exposure appears to be a key element of distinguishing the psychosocial from other approaches to social inequalities in health. This is particularly illustrated by the debates we briefly mentioned above, opposing proponents of the psychosocial model with the defenders of materialist approaches, be that from the Black Report or more recent research (Lynch et al. 2000). To materialists, social inequalities in health can be explained mainly by social conditions that result in direct exposure to pathogenic substances— pollution, harmful substances at work, etc.—, to degraded living conditions—unhealthy housing, malnutrition, and cold, —or even to “class cultures” that (involuntarily) induce “risk behaviors”, mainly consumption of unhealthy food, alcohol, and tobacco. Materialist approaches presupposes a gradualist analysis according to which a progressive increase in the standard of living—education, hygiene, etc.–leads mechanically to the disappearance of social inequalities in health.6 To those in favor of psychosocial theories, such as Michael Marmot, this assumption is invalidated by the existence of the social gradient in health, as revealed in the Whitehall I and II studies. The social gradient means that “the higher the social position, the better the health” (Marmot 2004). But most importantly, the gradient implies that health inequalities are not only measurable at the bottom, but throughout the social hierarchy. In this respect, the data confirms health status to be strongly correlated with stress factors determined by social status, such as the degree of control individuals have over their lives, how fulling their profession is to them and other factors that help to counteract the effects of stress, such as the number of friends,

6 The authors of the Black Report are in opposition to those who claim that in affluent societies material deprivation, and more generally the structural-materialist model, can no longer explain excess mortality among the working class. They refer to The Royal Commission, which has reported on the “remarkable stability of the unequal distribution of income over the past two decades” (Townsend et al. 1982, 115). They therefore conclude that social inequalities in health remain because material inequalities persist. One can deduce they assume social inequalities in health would cease to exist once a certain level of equity has been reached as regards the distribution of wealth. For more information on this debate see (Lynch et al. 2000; Marmot and Wilkinson 2001).

WHAT’S WRONG WITH THE BIOLOGIZATION OF SOCIAL …

91

symbolic rewards, and the quality of leisure activities which also depend on socioeconomic level (Marmot and Wilkinson 2001). By principally studying the socioeconomic conditions that directly result in the exposure of individuals to environments that are not conducive to health, materialists might fail to grasp that in rich societies, with a certain living standard, structural socioeconomic inequalities do not necessarily determine a state of absolute material poverty, with its accompanying exposure to risk factors and lack of resources. In such societies, structural conditions shape the unequal and relative distribution of symbolic privileges that are associated with socioeconomic status. These psychosocial factors can, in a certain instance, make individuals even more vulnerable to material factors of exposure. Consequently, the use of biology by proponents of the psychosocial model does not mean a retreat from the exploration of collective causes of health, but a development in the understanding of how structural causes lead to social inequalities in health. The combination of these factors is particularly relevant in rich societies where absolute poverty is no longer the only criterion of social inequality. What is more, the definition of social inequalities in health is transformed into a concept that no longer exclusively deals with the poorest populations in comparison to the richest, but all individuals according to the place they occupy in the social hierarchy—even those belonging to the most affluent classes. Within the limits of the case in hand, then, it would be exaggerated, if not simply inaccurate, to claim that biologization implies an individualization and a de-socialization of health problems in response to a liberal or neoliberal “regime of truth”. The thesis that the biologization of inequalities might lead social epidemiology to depoliticize the question of social inequalities is not only invalidated on the epistemic level. It is also invalidated by the moral economy of objectivity which frame social epidemiologists’ research program and their activism against social inequalities in health.

8

Conclusion

Those we met in our history of social epidemiology act as scientistsactivists by identifying the data and attesting to the fact that health is not only the responsibility of individuals, but also, and above all, the affair of politics. For these activist-scientists, emotional revolt conditions the

92

M. ARMINJON

norms of legitimate knowledge, both on scientific and political dimensions. In this regard, social epidemiologists collectively embrace a “moral economy of objectivity” that values the recognition of facts and the proper use of biology and statistical methodologies shared by the scientific community. This moral economy is based on the idea that the accumulation of evidence concerning social inequalities in health, as obvious as it is disparate (clinical, biological and epidemiological), provides the gain of objectivity necessary to induce in the general public and/or political authorities the same feeling of emotional outrage that they experience. Then, in theory, ad hoc public health strategies should follow. This makes it possible to reconsider where the main risk of depoliticization of the problems of social inequalities in health lies: the accumulation of data objectifying social problems is necessary but not sufficient for it to become a public and political concern. Thus, the main risk of depoliticization to which research on social inequalities in health is exposed does not seem to rest on objectification methods in general, whether they proceed from biologization. Data on social inequalities in health and the mechanisms that produce them have been accumulating for more than a century. Yet, while social crisis have historically stimulated interest in data objectifying social inequalities in health, these are more likely to be ignored when stability predominate (Webster 2002). Strategic invisibilization by governments are also documented, as Margaret Thatcher rejection of the Black Report (Berridge and Blume 2003). Other economic and social forces encouraging an individualistic and non-population-based public health intervention might also explain why objective data on social health inequalities are not more visible (Navarro 2009; Lynch 2017). It is thus more likely that the main risk of depoliticization lies on the social factors that tend to make them invisible and prevent them from becoming the object of a public debate.

References Acheson, Donald, and Great Britain, ed. 1998. Independent Inquiry into Inequalities in Health: Report. London: Stationery Office. Arminjon, Mathieu. 2016. Birth of the Allostatic Model: From Cannon’s Biocracy to Critical Physiology. Journal of the History of Biology 49: 397–423. https://doi.org/10.1007/s10739-015-9420-9.

WHAT’S WRONG WITH THE BIOLOGIZATION OF SOCIAL …

93

Arminjon, Mathieu. 2020a. The American Roots of Social Epidemiology and Its Transnational Circulation. Gesnerus 77: 35–43. https://doi.org/10.24894/ Gesn-en.2020.77002. Arminjon, Mathieu. 2020b. Rethinking the Normal and the Pathological. On Canguilhem’s Critical Physiology. In Vital Norms: Canguilhem’s The Normal and the Pathological in the Twenty-First Century, ed. Pierre-Olivier Méthot and Jonathan Sholl, 179–216. Paris: Hermann. Berkman, Lisa F., and Ichiro Kawachi. 2000. A Historical Framework for Social Epidemiology. In Social Epidemiology, ed. Lisa F. Berkman and Ichiro Kawachi, 3–12. New York: Oxford University Press. Berridge, Virginia, and Stuart Blume, ed. 2003. Poor Health: Social Inequality Before and After the Black Report. London: Frank Cass. Black, S. Douglas. 1980. Inequalities in Health: The Black Report. London: Department of Health and Social Security (DHSS). Bliss, Catherine. 2015. Defining Health Justice in the Postgenomic Era. In Postgenomics: Perspectives on Biology After the Genome, ed. Sarah S. Richardson and Hallam Stevens, 174–191. Durham: Duke University Press. Cannon, Walter B. 1916. Some Disorders Supposed to Have an Emotional Origin. New York Medical Journal 104: 870–873. Cannon, Walter B. 1922. Bodily Changes in Pain, Hunger, Fear and Rage. An Account of Recent Researches into the Function of Emotional Excitement. 2nd ed. New York and London: D. Appleton and Company. Cannon, Walter B. 1926. Physiological Regulation of Normal States: Some Tentative Postulates Concerning Biological Homeostatics. In A Charles Richet: ses amis, ses collègues, ses élèves, ed. Auguste Pettit, 91–93. Paris: Editions médicales. Cannon, Walter B. 1932. The Wisdom of the Body. New York: W. W. Norton. Cannon, Walter B. 1935. Stresses and Strains of Homeostasis. Journal of the Medical Sciences 189: 13–14. Cannon, Walter B. 1936. The Role of Emotion in Disease. Annals of Internal Medicine 9: 1453–1465. Cannon, Walter B. 2020 [1930] Walter B. Cannon. Conférences sur les émotions et l’homéostasie, Paris, 1930. Édition, introduction et notes par Mathieu Arminjon. Edited by Mathieu Arminjon. Lausanne: BHMS. Cannon, Walter B., Alfred Theodore Shohl, and W. S. Wright. 1911. Emotional Glycosuria. American Journal of Physiology—Legacy Content 29: 280–287. Cassel, John. 1974. Psychosocial Processes and “Stress”: Theoretical Formulation. International Journal of Health Services 4: 471–482. https://doi.org/ 10.2190/WF7X-Y1L0-BFKH-9QU2. Cassel, John. 1976. The Contribution of the Social Environment to Host Resistance. American Journal of Epidemiology 104: 107–123.

94

M. ARMINJON

Cassel, John, and Herman A. Tyroler. 1961. Epidemiological Studies of Culture Change: I. Health Status and Recency of Industrialization. Archives of Environmental Health 3: 25–33. https://doi.org/10.1080/00039896.1961.106 62969. Cassel, John, Ralph Patrick, and David Jenkins. 1960. Epidemiological Analysis of the Health Implications of Culture Change: A Conceptual Model. Annals of the New York Academy of Sciences 84: 938–949. Castagné, Raphaële, Valérie Garès, Maryam Karimi, Marc Chadeau-Hyam, Paolo Vineis, Cyrille Delpierre, Michelle Kelly-Irving, et al. 2018. Allostatic Load and Subsequent All-Cause Mortality: Which Biological Markers Drive the Relationship? Findings from a UK Birth Cohort. European Journal of Epidemiology 33: 441–458. https://doi.org/10.1007/s10654-018-0364-1. Cross, Stephen J., and William R. Albury. 1987. Walter B. Cannon, L. J. Henderson, and the Organic Analogy. Osiris 3: 165–192. CSDH. 2008. Closing the Gap in a Generation: Health Equity Through Action on the Social Determinants of Health: Final Report of the Commission on Social Determinants of Health. WA 525. Geneva: World Health Organization. Dubos, René. 1952. The White Plague: Tuberculosis, Man and Society. Boston: Little, Brown. Dubos, René. 1959. Mirage of Health: Utopias, Progress, and Biological Change. London: George Allen & Unwin. Dubos, René. 1963. Emerging Patterns of Disease. In The Voice of America Forum, 1–11. Man Under Stress Series 9. Washington, DC: US Agency. Dubos, René. 1965. Man Adapting. Silliman Memorial Lectures. New Haven and London: Yale University Press. Dubos, René. 1974. Homeostasis, Illness, and biological Creativity. Lahey Clinic Foundation Bulletin 23: 94–100. Dubos, René, Dwayne Savage, and Russell Schaedler. 2005 [1966]. Biological Freudianism. Lasting Effects of Early Environmental Influences. International Journal of Epidemiology 34: 5–12. https://doi.org/10.1093/ije/dyh309. Dupras, Charles, and Vardit Ravitsky. 2016. Epigenetics in the Neoliberal “Regime of Truth”: A Biopolitical Perspective on Knowledge Translation. The Hastings Center Report 46: 26–35. https://doi.org/10.1002/hast.522. Epstein, Frederick H., and Ronald D. Eckoff. 1967. The Epidemiology of High Blood Pressure-Geographic Distributions and Etiological Factors. In The Epidemiology of Hypertension, ed. Jeremiah Stamler and Rose Stamler, 155–166. New York: Grune and Stratton. Eyer, Joseph, and Peter Sterling. 1977. Stress-Related Mortality and Social Organization. The Review of Radical Political Economics 9: 1–44. Fassin, Didier. 2009. Les économies morales revisitées. Annales. Histoire, Sciences Sociales 64. Cambridge University Press: 1237–1266. JSTOR.

WHAT’S WRONG WITH THE BIOLOGIZATION OF SOCIAL …

95

Fausto-Sterling, Anne. 2004. Refashioning Race: DNA and the Politics of Health Care. Differences: A Journal of Feminist Cultural Studies 15: 1–37. Fausto-Sterling, Anne. 2005. The Bare Bones of Sex: Part 1—Sex and Gender. Signs 40: 1491–1527. Geiger, H. Jack. 2002. Community-Oriented Primary Care: A Path to Community Development. American Journal of Public Health 92: 1713–1716. Haraway, Donna J. 1991. Simians, Cyborgs, and Women: The Reinvention of Nature. New York: Routledge. Jackson, Mark. 2013. The Age of Stress: Science and the Search for Stability. Oxford: Oxford University Press. Krieger, Nancy. 2001. Theories for Social Epidemiology in the 21st Century: An Ecosocial Perspective. International Journal of Epidemiology 30: 668–677. Kuznick, Peter. 1987. Beyond the Laboratory: Scientists as Political Activists in 1930s America. Chicago: The University of Chicago Press. Lock, Margaret. 2013. The Epigenome and Nature/Nurture Reunification: A Challenge for Anthropology. Medical Anthropology 32: 291–308. https:// doi.org/10.1080/01459740.2012.746973. Lorraine Daston. 1995. The Moral Economy of Science. Osiris 10: 2–24. Louvel, Séverine, and Alexandra Soulier. 2022. Biological Embedding vs. Embodiment of Social Experiences: How These Two Concepts Form Distinct Thought Styles Around the Social Production of Health Inequalities. Social Science & Medicine 314: 115470. https://doi.org/10.1016/j.socscimed. 2022.115470. Lynch, John W., George Davey Smith, George A. Kaplan, and James S. House. 2000. Income Inequality and Mortality: Importance to Health of Individual Income, Psychosocial Environment, or Material Conditions. British Medical Journal 320: 1200–1204. Lynch, Julia. 2017. Reframing Inequality? The Health Inequalities Turn as a Dangerous Frame Shift. Journal of Public Health 39: 653–660. https://doi. org/10.1093/pubmed/fdw140. Macintyre, Sally. 1997. The Black Report and Beyond What Are the Issues? Social Science & Medicine 44. Health Inequalities in Modern Societies and Beyond: 723–745. https://doi.org/10.1016/S0277-9536(96)00183-9. Marmot, Michael G. 2001. From Black to Acheson: Two Decades of Concern with Inequalities in Health. A Celebration of the 90th Birthday of Professor Jerry Morris. International Journal of Epidemiology 30: 1165–1171. https:// doi.org/10.1093/ije/30.5.1165. Marmot, Michael G. 2003. Understanding Social Inequalities in Health. Perspectives in Biology and Medicine 46. Johns Hopkins University Press: S9–S23. https://doi.org/10.1353/pbm.2003.0070. Marmot, Michael G. 2004. Status Syndrome: How Your Social Standing Directly Affects Your Health and Life Expectancy. London: Bloomsbury.

96

M. ARMINJON

Marmot, Michael. G., Geoffrey Rose, Martin Shipley, and Peter J. Hamilton. 1978. Employment Grade and Coronary Heart Disease in British Civil Servants. Journal of Epidemiology & Community Health 32: 244–249. https://doi.org/10.1136/jech.32.4.244 Marmot, Michael G., Martin J. Shipley, and Geoffrey Rose. 1984. Inequalities in Death—Specific Explanations of a General Pattern? The Lancet 323. Originally Published as Volume 1, Issue 8384: 1003–1006. https://doi.org/10.1016/ S0140-6736(84)92337-7. Marmot, Michael G., and Richard G. Wilkinson. 2001. Psychosocial and Material Pathways in the Relation Between Income and Health: A Response to Lynch et al. BMJ: British Medical Journal 322: 1233–1236. Marmot, Michael G., and S. L. Syme. 1976. Acculturation and Coronary Heart Disease in Japanese-Americans. American Journal of Epidemiology 104: 225– 247. McEwen, Bruce S., and Eliot Stellar. 1993. Stress and the Individual: Mechanisms Leading to Disease. Archives of Internal Medicine 153: 2093–2101. McEwen, Bruce S., and John C. Wingfield. 2003. The Concept of Allostasis in Biology and Biomedicine. Hormones and Behavior 43: 2–15. https://doi. org/10.1016/S0018-506X(02)00024-7. McKeown, Thomas, and Robert G. Brown. 1955. Medical Evidence Related to English Population Changes in the Eighteenth Century. Population Studies 9: 119. https://doi.org/10.2307/2172162. Meloni, Maurizio. 2015. Epigenetics for the Social Sciences: Justice, Embodiment, and Inheritance in the Postgenomic Age. New Genetics and Society 34: 125–151. https://doi.org/10.1080/14636778.2015.1034850. Moore, Kelly. 2013. Disrupting Science: Social Movements, American Scientists, and the Politics of the Military, 1945–1975. Princeton: Princeton University Press. Navarro, Vicente. 2009. What We Mean by Social Determinants of Health. International Journal of Health Services 39: 423–441. https://doi.org/10.2190/ HS.39.3.a. Neser, William B., Herman A. Tyroler, and John C. Cassel. 1971. Social Disorganization and Stroke Mortality in the Black Population of the North California. American Journal of Epidemiology 93: 166–175. Niewöhner, Jörg. 2011. Epigenetics: Embedded Bodies and the Molecularisation of Biography and Milieu. BioSocieties 6: 279–298. https://doi.org/10.1057/ biosoc.2011.4. Oppenheimer, Gerald M., Ronald Bayer, and James Colgrove. 2002. Health and Human Rights: Old Wine in New Bottles. Journal of Law, Medicine & Ethics 30: 522.

WHAT’S WRONG WITH THE BIOLOGIZATION OF SOCIAL …

97

Porter, Dorothy. 2011. Health Citizenship: Essays in Social Medicine and Biomedical Politics. Perspectives in Medical Humanities. Berkeley: University of California, Medical Humanities Consortium. Prior, Lucy, David Manley, and Kelvyn Jones. 2018. Stressed Out? An Investigation of Whether Allostatic Load Mediates Associations Between Neighbourhood Deprivation and Health. Health & Place 52: 25–33. https://doi.org/ 10.1016/j.healthplace.2018.05.003. Richet, Charles. 1919. L’homme stupide. Paris: Flammarion. Rose, Nikolas. 2001. The Politics of Life Itself. Theory, Culture & Society 18. Sage: 1–30. https://doi.org/10.1177/02632760122052020. Scott, James C. 2006 [1976]. The Moral Economy of the Peasant: Rebellion and Subsistence in Southeast Asia. New Haven: Yale University Press. https://doi. org/10.12987/9780300185553. Shostak, Sara, and Margot Moinester. 2015. The Missing Piece of the Puzzle? Measuring the Environment in the Postgenomic Moment. In Postgenomics: Perspectives on Biology After the Genome, ed. Sarah S. Richardson and Hallam Stevens, 192–209. Durham: Duke University Press. Steptoe, Andrew, and Michael G. Marmot. 2002. The Role of Psychobiological Pathways in Socio-Economic Inequalities in Cardiovascular Disease Risk. European Heart Journal 23: 13–25. https://doi.org/10.1053/euhj.2001. 2611. Sterling, Peter. 2004. Principles of Allostasis: Optimal Design, Predictive Regulation, Pathophysiology, and Rational Therapeutics. In Allostasis, Homeostasis, and the Cost of Physiological Adaptation, ed. Jay Schulkin, 17–64. Cambridge, MA: Cambridge University Press. Sterling, Peter, and Joseph Eyer. 2005 [1988]. Allostasis: A New Paradigm to Explain Arousal Pathology. In Handbook of Life Stress, Cognition and Health, Shirley Fisher and James Reason, 629–649. Chichester: Wiley. Syme, S. Leonard. 2005. Historical Perspective: The Social Determinants of Disease—Some Roots of the Movement. Epidemiologic Perspectives & Innovations 2: 2–7. https://doi.org/10.1186/1742-5573-2-2. Syme, S. Leonard, Merton M. Hyman, and Philip E. Enterline. 1965. Cultural Mobility and the Occurrence of Coronary Heart Disease. Journal of Health and Human Behavior 6: 178. https://doi.org/10.2307/2948634. Syme, S. Leonard, Thomas W. Oakes, Gary D. Friedman, Robert Feldman, Abraham B. Siegelaub, and Morris F. Collen. 1974. Social Class and Racial Differences in Blood Pressure. American Journal of Public Health 64: 619–620. https://doi.org/10.2105/AJPH.64.6.619. Thompson, Edward P. 1966. The Making of the English Working Class. New York: Vintage Books. Tournès, Ludovic. 2011. Sciences de l’homme et politique. Les fondations philanthropiques américaines en France au XXe siècle. Classiques Garnier.

98

M. ARMINJON

Townsend, Peter B., Margaret Whitehead, and Nick Davidson. 1982. Inequalities in Health: The Black Report. London: Penguin Books Ltd. Vineis, Paolo, Cyrille Delpierre, Raphaële Castagné, Giovanni Fiorito, Cathal McCrory, Mika Kivimaki, Silvia Stringhini, Cristian Carmeli, and Michelle Kelly-Irving. 2020. Health Inequalities: Embodied Evidence Across Biological Layers. Social Science & Medicine 246: 112781. https://doi.org/10.1016/j. socscimed.2019.112781. Vineis, Paolo, Mauricio Avendano-Pabon, Henrique Barros, Mel Bartley, Cristian Carmeli, Luca Carra, Marc Chadeau-Hyam, et al. 2020. Special Report: The Biology of Inequalities in Health: The Lifepath Consortium. Frontiers in Public Health 8: 118. https://doi.org/10.3389/fpubh.2020.00118. Webster, Charles. 2002. Investigating Inequalities in Health Before Black. Contemporary British History 16: 81–104. https://doi.org/10.1080/713 999462. West, Patrick. 1998. Perspectives on Health Inequalities: The Need for a Lifecourse Approach. Medical Research Council Social & Public Health Sciences Unit: 1–22. Wilkinson, Richard. 1976. Dear David Ennals. New Society: 567–568.

Integration in Environmental Health and Exposome Research: Epistemological Issues

Which Integration for Health? Comparing Integrative Approaches for Epidemiology Stefano Canali

1

Introduction

Biomedical research and clinical care are constantly filled with new promises, trends, and keywords that frame health and disease in new and different terms. These keywords intersect at various levels, appearing in biomedical research and funding schemes, in policy discussions and political speeches, as well as in the media and public discourses. Some of these concepts often end up having a short life span, but others stay and sometimes get entrenched within specific research communities or even ways in which health policy is conducted. The relevance and importance of these concepts and their epistemological consequences merit our attention—in particular if we want to make sense, conceptually, of what these mean and promise and when they have shaped lines of research and can further do so in the future. In philosophy of medicine, significant work has focused on

S. Canali (B) Department of Electronics, Information and Bioengineering & META—Social Sciences and Humanities for Science and Technology, Politecnico Di Milano, Milan, Italy e-mail: [email protected]

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 É. Giroux et al. (eds.), Integrative Approaches in Environmental Health and Exposome Research, https://doi.org/10.1007/978-3-031-28432-8_5

101

102

S. CANALI

the merits and pitfalls of approaches and notions including evidence-based medicine (Worrall 2002, 2007; Stegenga et al. 2017), personalised and precision medicine (Vogt et al. 2016; Prainsack 2020; Plutynski 2022), genomics and postgenomics (Richardson and Stevens 2015; Hilgartner 2017; Gibbon et al. 2020). Less attention has instead been given to conceptualisations of health and disease in the context of epidemiology and public health, but the area is ripe with new and different ways of discussing and investigating health and disease, from new types of collection and use of health data to the new methodological and statistical treatment of health data, from new conceptualisations of health to the employment of new types of causal inference. In this chapter, I analyse a specific type of concepts: concepts aimed at framing the relations between the health of populations and the environment. As we will see, the focus on these relations is a recent shift in epidemiology, but one that has gained increasing momentum in recent years. I will specify my discussion on the following three concepts: the exposome, a notion aimed at capturing all the exposures that individuals and populations experience; planetary health, a conceptualisation of the impact of local and multispecies environments on human health; and global health, a concept aimed at framing the health of populations as the result of different social factors of different and diverse populations. In analysing these new concepts, I will apply a specific methodological focus by looking at data integration, as the ensemble of data practices aimed at using different types and bodies of data for the study of specific phenomena. This focus is grounded on recent philosophical literature on the epistemology of integration in the life and health sciences and the epistemology of scientific data. As I will show, this focus enables me to identify significant assumptions and commitments, merits, and limitations of the three concepts. The chapter is structured as follows. I start by motivating my focus on a specific type of integration—data integration—in relation to ongoing research in philosophy of science and medicine (Sect. 2). I then use this focus as a conceptual lens to analyse and discuss three crucial keywords that frame health and biomedical research in specific ways and, as I argue, are distinct approaches to the integration of different types of data. On this basis, I focus on the exposome (Sect. 3), global health (Sect. 4), and planetary health (Sect. 5), critically analysing their approaches and identifying limits their limitation.

WHICH INTEGRATION FOR HEALTH? COMPARING …

2

103

The Many Faces of Integration: Why Data Integration Matters

Integration is a significant focus of recent philosophy of science, particularly in the context of discussions on the metaphysical and epistemological foundations of the life sciences. These have traditionally been discussed as particularly fragmented, with highly diverse communities, approaches, methodologies, theories, styles of explanation, commitments, and goals. In this sense, the analysis of these aspects of the life sciences has been intertwined with questions of reduction, unification, and indeed integration, with a focus on the extent to which elements of biological theory or ontology can be integrated and reduced to physics, the plurality and diversity of the life sciences is something that can and should be integrated and reduced, or should lead to unifications at the conceptual or methodological level (Dupré 1996). Several philosophers of biology have argued that different theories in biology cannot be unified nor reduced to more fundamental ones (Mitchell and Dietrich 2006) and the use of diverse methods, models, and representations is crucial in a number of areas of the life sciences (Mitchell and Gronenborn 2017), as many questions and problems require explanations developed in different biological disciplines and with different scientific aims (Brigandt 2010). The question of integration thus brings substantial results from the philosophical literature and can be analysed from several different points of view in the context of contemporary life and health sciences, and epidemiology in particular (Giroux et al. 2021). In the context of this chapter, I will focus on a specific type of integration: data integration. Data integration can be defined as the set of data practices involved in making different types of data usable as single bodies of evidence. For example, data integration can entail working on samples collected at different points of data collection, for instance blood samples collected in a longitudinal study, to make sure that all the samples can be compared and used as evidence for a broader study, for instance a study of cardiovascular disease in the general population. At the same time, as a result of the aforementioned fragmentation of the life and health sciences, in this context the notion of data can refer to many different objects, usually comes in significantly different types, and is collected by different communities, for different purposes and with diverse commitments. In addition, the life and health sciences present an interesting case in which the standardised production of large volumes of data has often been mentioned

104

S. CANALI

as a potential game changer (Golub 2010; Weinberg 2010), but practices of data handling, storing, and analysis have strong continuity with longstanding approaches (Müller-Wille and Charmantier 2012; Leonelli 2016; Strasser 2019). This entails that substantial work needs to be conducted on preparing diverse data sets for data analysis, in order to compare different findings, building a more robust evidential basis for a specific claim, contrasting quality from various research settings. Data also sit at the crossroads of many current trends of the field, especially in the biomedical sciences. For example, personalised and precision medicine, as the attempts to take into account individual variables into the study and prevention of disease, are largely built on the assumption that the use of large data sets “can account for an increasing number of factors that influence health and disease, and that these data can be used to stratify the population and health problems according to various characteristics” (Green and Vogt 2016). Similarly, the molecularisation of the health sciences (Boniolo and Nathan 2017) and postgenomics (Richardson and Stevens 2015) are largely based on the use and integration of new data sets, at different levels of abstraction and at increasing volumes. These features suggest that data integration involves technical considerations about the intrinsic properties of the data, such as their interoperability, quality, resolution, etc. However, results from the philosophy of scientific data have shown that data integration also involves significant inferences, assumptions, strategies that provide a significant window into the epistemology of biomedical research and data (Leonelli 2013). In this sense, focusing on data integration is an important analytical choice for philosophy of science, but is particularly interesting in the context of epidemiology, health, and especially the approaches I discuss in this chapter. Epidemiological research is traditionally concerned with diverse and large types of data (Morabia 2004), but the recent raise in the volume, diversity, heterogeneity of data in the sciences (Leonelli 2016; Leonelli and Tempini 2020) has often been discussed as a substantial novelty for the field, potentially opening new and uncharted territories for both research and policy-making (Holmberg et al. 2013; Hogle 2016; Fleming et al. 2017). For example, the increasing availability of new data on the environment and climate is opening new opportunities for the study of the impact of different types of environments on population health and shaping the ways in which the environment is conceptualised and operationalised across approaches in epidemiology (Canali and Leonelli 2022).

WHICH INTEGRATION FOR HEALTH? COMPARING …

105

At the same time, the availability of new sources of data on population behaviour, for instance through social media, personal health applications, and monitoring devices, is supposed to create new disciplines and areas of research in the life sciences, such as digital epidemiology, relying on the increasing datafication of health and medicalisation of population activities through data coming from outside epidemiology and scientific research more generally (Salathé 2018; Mittelstadt et al. 2018; Klingwort and Schnell 2020). Data integration thus gives us a window into the approaches, assumptions, and conceptualisations that are crucial aspects of the epistemology of contemporary epidemiology (Giroux, this volume). It is on this basis that in the chapter I focus on the integration of data as an analytical lens to discuss the different approaches to integration, identify assumptions and commitments, and discuss conceptual gaps and methodological issues. As we will see, the approaches that I discuss in the chapter are all aimed at a certain type of integration at the level of health and environment, which elicits questions on which type of integration they aim for, the extent to which such integration is actually achievable or achieved, and what this implies more generally. In particular, the concept of the exposome is a concept that aims to integrate different dimensions and categories of exposure; the concept of planetary health attempts to integrate different dimensions of the health of populations and environments; and the concept of global health is a way of integrating different types and sources of health at a global level. In the context of this chapter, I will approach them with the methodological choice of focusing on data integration and discussing the different types of data integration that are part of these approaches to environmental research. There are clearly various other aspects of integration that merit discussion in the context of these three approaches and various other chapters of this book go in that direction (see Meloni, Russo, Giroux, this volume). My choice of focusing on data integration builds on this work and aims to identify epistemic assumptions, features, and issues that are involved in these three approaches and emerge at the level of data practices. Framing these concepts as data integration approaches is a particular way of studying them and specifying philosophical questions of integration with a focus on an increasingly important area of scientific practice. This is a specific methodological choice that frames and specifies the work I do in the chapter. Other concepts, frameworks, approaches could be chosen to discuss integration and focus on data integration, including for instance social determinants of health and disease, precision medicine,

106

S. CANALI

toxicology. Yet in this chapter I focus on attempts at integration between health and environment in the study of population health and look at their relations with current research projects and data practices that have been largely successful at securing funding and establishing research centres in the last few years. As I will discuss at several points of the remainder of the chapter, while I do not aim for a broad generalisation of my findings, the approaches I chose to focus on also share significant features with the aforementioned and other approaches in contemporary epidemiology.

3 The Exposome: Integrating Molecular and Environmental Data The exposome has raised as a specific approach, concept, and area of research in the last decade in epidemiology and beyond. The exposome was first conceived as a specific notion, which conceptually would entail a new expansion of the traditional notion of exposure by adding more specific dimensions at the external level and a new dimension at the internal level (Wild 2005). The rationale here was that epidemiologists needed to expand and specify the ways in which exposure and the environment were considered and different types of data were collected; a crucial element of inspiration was genomics and the Human Genomic Project, both in the sense of providing molecular tools for data collection and analysis and in the sense of adding the new dimensions to internal exposure. As such, conceptually the exposome is presented as the totality of all exposures that individuals experience in a specific point in time and cumulatively though all their life and methodologically is made of three types and categories of exposure: general external exposure, specific external exposure, and internal exposure (Wild 2012). The idea here is that individuals and populations can be exposed to environmental agents at different levels, both internally and externally: for instance at the general external level with processes connected to socio-economic status, at a specific external level with processes such as specific chemical pollutants and infectious agents, or at the internal level with processes including endogenous responses to external exposure (e.g. inflammation, oxidative stress, etc.). In this sense, conceptually the exposome has been framed as a counterpart to the genome—i.e., taking inspiration from the methodological and conceptual status of the genome while intending to critically complement genomic research with the study of health and

WHICH INTEGRATION FOR HEALTH? COMPARING …

107

disease with epidemiological tools and approaches, in particular individual and cohort studies. The exposome as an area of research has indeed relied extensively on molecular data collection and analytic tools, by integrating omics techniques into epidemiology (Boniolo and Nathan 2017). Omics are molecular techniques that in the genomic context are used to study molecules and processes within a cell, including for example metabolism and protein development and quantify and collect data on these processes. In the exposome context, omics data have been used to analyse the internal dimension of exposure and develop exposure profiles, which measure processes and entities at the internal level and can be analysed in comparison and contrast with external exposure data (Russo and Vineis 2016). For instance, omics techniques can be used to analyse blood samples collected in longitudinal studies and measure adducts at the molecular and internal level whose presence can be connected to external toxicants from e.g. pollution. The resulting exposure profiles can be contrasted and compared with quantitative data on the presence of these pollution toxicants and thus data on external exposure at different levels, as a way to study the influence on pollution on the internal molecular environment and more generally the effects of pollution on the development of health and disease. The extension of these genomic tools and methods into epidemiology has contributed to the raising influence of the exposome in contemporary epidemiology (Canali 2020a). As such, research on the exposome has received substantial and specific funding, for instance with dedicated programmes in the EU, and several research centres and units have been established, for example in the US. These processes have contributed to establishing the exposome as a way of doing epidemiology, based on the integration of different types of exposures, that are measured and quantified through molecular and data-intensive tools.1 As a result, the exposome approach is an important point of focus for philosophical, methodological, historical analyses of epidemiology 1 These features have led to significant differences between traditional approaches to exposure in epidemiology and the exposome, for instance at the level of the conceptualisation of internal exposure, the use of omics and other molecular tools, and related types of funding required and sought after. Yet exposome research has been only partially successful at securing long-term funding and, as helpfully suggested by one anonymous reviewer, the exposome approach has not taken over epidemiology in general (Canali 2020a).

108

S. CANALI

(see e.g. Illari and Russo 2016; Giroux et al. 2021; Russo et al. in this volume). In this sense, in the context of this chapter, I frame the exposome as a significant and specific approach to the integration of data in epidemiology. Following Giroux (this volume), there are several different dimensions of integration that need to be discussed in the case of the exposome—data integration is a specific and particularly interesting type in this context. As I have mentioned, the data integration that takes place in exposome research is largely based on the extension of genomic solutions and techniques to epidemiology. Practically, this mostly means integration between molecular data on the internal dimension of exposure and environmental data on various types of external exposure, which frames the use of genomic data as a platform to link biological and environmental data to study exposure. What count as biological and environmental data can vary rather substantially depending on the specific focus of research groups and projects. In the context of exposome research, biological data have mostly been identified with molecular and omics data, while environmental data have varied from climatic data and environmental sampling to the collection of questionnaires data in cohort studies and social scientific data more generally.2 This has led to the development of new epistemic strategies for the integration of biological and environmental data in exposome research and their alignment with existing methodological approaches in epidemiology (Canali 2020b). In particular, the use of omics and molecular data has pushed for a microscopic focus on internal exposure as a central aspect of epidemiological research—to the point that this focus has led to changes also in the way external exposure is studied and environmental data are collected. In general terms, the focus has led to a push for more standardisation and quantification of the presence and effects of environmental exposure. For instance, exposome researchers have started to collaborate with disciplines in computer and information science, such as information systems, to quantify and estimate the presence of specific chemicals and pullulations in the external

2 This variety has consequences for the conceptual dimensions and boundaries of crucial

notions for epidemiology, such as exposure and environment. As one anonymous reviewer helpfully suggested, the concept of environment is often quite vague in the context of exposome research: see Canali & Leonelli (2022) for more on the concept of the environment in exposome research and data-intensive epidemiology more generally and its relation with the availability of new environmental data.

WHICH INTEGRATION FOR HEALTH? COMPARING …

109

environment. On this basis, for example, the individual concentration of chemicals connected to air pollution, such as particulate air matter, is available to be compared and contrasted with molecular data on the presence of chemicals in the body and reactions that can be connected to health and disease (see e.g. Gulliver et al. 2018).3 The focus on molecular and internal components of exposure and the push for standardisation and quantification at the external level of exposure is not surprising. Part of the rationale and one of the aims of the original introduction of the exposome was the need for more standardisation in environmental sampling and the idea of transferring genomic solutions in epidemiology was indeed to match the level of quantification and precision of genomics for epidemiological and environmental research (Wild 2005; Rappaport and Smith 2010). The need for a movement in this direction was elicited by a general lack of focus on the impact of the environment of changes in health and disease in populations, but also the scattered and irregular state of environmental sampling and difficulties with data integration and collaboration across the health and environmental sciences.4 The approach applied by the exposome is thus based on the use of molecular data as a platform to use environmental and particularly climatic data in the study of population health. In this direction, exposure profiles developed through omics techniques are the basis for the collection of data on external exposure of a similar, comparable level of abstraction and resolution. For instance, this has led to the collection of geographic data on the presence of pollutants and toxicants in the environment surrounding participants to a cohort studies, such as data on particulate air matter and other air pollutants (Canali 2020b). In these ways, one of the epistemic goals of the exposome approach has been to analyse correlations between molecular and environmental data with

3 As a result, in exposome research environmental data are often equated with climatic and geographical data. This is not just a feature of exposome research, as similar considerations can be traced to other data-intensive approaches in epidemiology as well as other areas of the life and health sciences, such as toxicology, exposure science, biomarkers research (Canali and Leonelli 2022). 4 This is connected to more general aspects of the history and development of the health sciences, epidemiology in particular, and the environmental sciences. For more on this in the context of history of notion of environment and research in the environmental sciences, see Warde et al. (2018).

110

S. CANALI

the aim of identifying causal relations between health, social and environmental features and study their effects on health. In this direction, in the exposome context, several statistical methods and approaches have been developed and used, such as the “meet-in-the-middle approach” (Chadeau-Hyam et al. 2011, 2013). These have been interpreted philosophically as ways to move beyond the sole focus on associations—a typical feature of epidemiological research—and collect elements of mechanistic evidence, which together with difference-making evidence are necessary to develop robust causal claims in the health sciences (Russo and Williamson 2007; Russo 2009; Illari and Russo 2016; Russo and Vineis 2016). These have clearly been steps in a new and promising direction, but this approach to data integration also suffers from significant limitations— as exposome researchers are aware. The reliance on genomic solutions and the use of molecular and omics data as proxies to study environmental features has been considered an extension of the reductionism of genomics to new areas of the health sciences (Landecker and Panofsky 2013; Shostak and Moinester 2015). Using, for instance, inflammation and oxidative stress at the molecular data as proxies for the impact of socio-economic and other environmental conditions can be seen as a problematic way of quantifying qualitative processes. As a response to this criticism, it is important to note that the data integrated through the exposome approach can be used to develop mechanistic explanations of the impact of the environment on health and disease (Illari and Russo 2016; Russo and Vineis 2016). Yet using molecular data to build mechanistic evidence faces challenges as well. Omics techniques are used in exposome research to look for specific entities whose presence may be due to and depend on external exposure, rather than the activities and dynamics that can make up mechanistic explanations—to the point that exposome researcher are usually extremely cautious when it comes to mechanistic explanations and, more generally, causal claims on the basis of molecular data (Canali 2019). In addition, more recent exposome projects have relied on genomic data as a potential proxy for the study of some social features and determinants of health and disease (Ghiara and Russo 2019). The collection and development of mechanistic evidence about the pathways through which social and environmental features can have an impact on health and disease is crucial for the ability to act on those pathways and improve health for a population. Yet, often, mechanistic evidence is not sufficient here: many environmental and social

WHICH INTEGRATION FOR HEALTH? COMPARING …

111

dynamics that have a significant impact on a population operate through various different pathways and mechanisms. For instance, racism and poverty can operate in various different ways and may be sources of exposure at the level of education, pollution, diet, mentorship, etc. (Valles 2019). Knowing about one or more of these is crucial, but in many cases it can be more important to know the differences and dependencies that these causes have with disease and act on them, instead of getting precise pictures of how these have an impact (Giroux, this volume; Valles 2021).

4

Planetary Health: Integrating Diverse Environmental Data

The need to expand the integration of data on the environment in epidemiology is a central goal and defining feature of the exposome approach. Yet the exposome as an approach and area of research is not alone in the landscape of contemporary epidemiology—various approaches and new conceptualisations aim to capture the needs and types of integration of the environment for health. Planetary health is one of these novelties and is an increasingly important concept for epidemiology and the life sciences more generally.5 Planetary health has been introduced in the last decade as a conceptual and methodological expansion for the field of public health. As an approach, planetary health has encouraged an inclusion of more and different human, physical, animal, climate environments as the types of environments that should be analysed in the study of population health. As discussed by Richard Horton and colleagues in the introduction of the concept in 2014, planetary health is supposed to direct more light on the “fragility of our planet and our obligation to safeguard the physical and human environments” (Horton et al. 2014, 847). Planetary health is thus based on a twofold realisation: health and environment are intrinsically connected phenomena, as changes and features of different environments have an impact on health at various levels; conversely, the environment is clearly shaped by changes and attitudes of populations related to health. As a result, in population health the state of health and disease of environments and populations is equated, with the goals of looking at health

5 In turn, planetary health shares significant features with and is close to other conceptualisations, such as “one health” and “spatial epidemiology”.

112

S. CANALI

as a complex, multi-level, but essentially unitary phenomenon. Health is seen as a multi-level property precisely because of the interactions and interrelations between environments, species, populations, and individuals. But it is unitary in the sense that health is a property, according to the planetary point of view, of both individuals and populations on the one hand and environments and ecosystems on the other. Hence the need to consider these complex phenomena and study them together, for the improvement of both. Since the introduction of the notion, planetary health has gained momentum with the expansion of new declarations and campaigns, such as the São Paulo Declaration on Planetary Health from the fall of 2021 (Myers et al. 2021), and has now dedicated journals, including for instance a speciality journal published by The Lancet.6 This shows a first, significant difference between planetary health and conceptualisations of health and disease of the exposome. Planetary health has been first and foremost a contribution for political discussions at the level of policy on the environment, biodiversity, social determinants of health and, as such, consequently on the management, organisation, and funding on these issues. This means that planetary health is currently a conceptual framework and political project, rather than a very concrete methodological approach such as the exposome. This does not mean, however, that the concept should be dismissed as something that is significant “only” at the policy level and is thus not particularly interesting for discussions on the epistemology and methodology of the health sciences. On the contrary, conceptualisations, values, and decisions at the political and economic level of the sciences clearly have a significant epistemic impact on the ways in which research is funded, managed, directed but also conducted at a concrete and practical level. This is why the focus on data integration can be helpful to analyse and discuss the assumptions, goals, and limits of an approach and conceptualisation such as planetary health. The approach to data integration for health applied by planetary health is based on an expansion of the types of data that are considered and used as environmental data, with a push for more focus on different environments and the interactions between different species that inhabit 6 See The Lancet Planetary Health: https://www.thelancet.com/journals/lanplh/home (accessed October 2022). As I write this, Richard Horton, the author of the paper introducing the concept of planetary health (Horton et al. 2014) is editor in chief of The Lancet.

WHICH INTEGRATION FOR HEALTH? COMPARING …

113

these environments. This might seem like a clearly important focus and almost trivial realisation, but is already quite different from the approach of exposome research, where the expansion is mostly focused on introducing genomic techniques and data in epidemiology and thus integrating population health and the environment on the basis of an expansion of internal exposure and health data. As we have seen with the discussion of the exposome, studies of population health based on environmental data are far from abundant in both the environmental and health sciences. On the one hand, research on environmental changes and more generally the idea that the environment is something that changes and should be studied for health is relatively recent. Before the mid-1800s, the notion of “environment” referred to the set of background circumstances, conditions, and stimuli that shape the character of an individual, situation, or phenomenon (Warde et al. 2018). According to this meaning of the notion, the environment was always the environment of something; and the focus was on this something, i.e. the individual, situation or phenomenon affected by the environment. While the notion also referred to conditions in the environment that needed to be acted upon, these conditions were mostly considered unchangeable and immutable per se— the focus of possible changes was again on the subject, rather than on the environment itself. In epidemiology, this focus changed with the hygiene and sanitary movements of the second half of the nineteenth century, as agents were identified as causes of infectious disease and loci of intervention beyond the subject. Historically, the development and success of epidemiology is deeply connected to these policy interventions, but the study of population health has long focused its priority on individual behaviors and lifestyle choices, and thus not necessarily on the environment (Morabia 2004). At the same time, in the natural and physical sciences, the shift towards viewing the environment as a subject of specific interest and agency only happened in the second half of the twentieth century, in connection with various other trends in the sciences, politics, and society more generally. As a result, the study of the environment and population health have interacted relatively rarely. This background on the interactions between the study of the environment and population health has concrete consequences on the limitations of the planetary health approach to data integration. The studies and methods used to analyse the relations between population health and its environment in epidemiology are configured in ways that allow for the integration of exposure data from different sources and their assemblage

114

S. CANALI

in specific configurations to study health and disease conditions of interest (Bauer 2013). These approaches to the collection and analysis of environmental data enable epidemiologists to study populations over extended periods of time and compare them over different configurations through “probabilistic thinking” (Morabia 2004). As a result, these methods are usually coordinated and centralised efforts that generate observational data about the environment, but in ways that tend to be quite different and largely independent from the experiments, models, and simulations employed in environmental research. Therefore, in epidemiology the study of the environment in relation to population health is normally based on data that are highly related to a population and environmental exposure and only partially to the direct study of the environment. In most cases, no direct sampling of environmental pollutants can be conducted and data about the internal biochemical environment are used as a proxy for variables tracking specific features and changes in the external environment. A more direct focus on the environment for population health is thus made difficult by differences between the types of data collected for environmental and health research, their varying time scales, and frequencies (Fleming et al. 2017). In addition, epidemiological methods usually employ regression approaches that focus on few exposure factors for small groups and single health outcomes, which is problematic when trying to address the overall impact of the environment on health. While epidemiologists are clearly interested in monitoring and surveilling population health, climate and environmental scientists have developed tools and methods for the estimate and prediction of climate events (Parker 2018). This is an issue for planetary health research, particularly at the level of gaps in the collection of data that are considered crucial for the study of health and disease in their connection with diverse types of environments. Similar gaps and issues are connected to disciplinary boundaries between epidemiology and environmental research, which make collaborations and interdisciplinarity difficult in this context. The epistemic impact of these issues, moreover, is particularly severe at the level of data interpretation and use as evidence, particularly when it comes to causal analysis and inference. The aforementioned gaps in the causal attribution of environmental determinants of health and disease are more severe in the case of missing and unclear data collection on the environment.

WHICH INTEGRATION FOR HEALTH? COMPARING …

5

115

Global Health: Integrating Diverse Health Data

With both the introduction of the exposome and planetary health concepts, we see new and different approaches to the integration of health and environmental data to study the relation between health and the environment. In this context, an additional approach that is important to discuss from a data integration perspective is global health. Global health is the consideration of health in global terms: a view of health and disease as the result of the differently and unequally distributed needs and issues of the various population that live in different parts of the world and in related environments.7 The concept emerged in the late 1990s as a way of expanding the notion of international health that was extensively used until then with reference to the issue of epidemics and their control beyond the borders of individual nation-states. Global health was introduced in this context as a way of expressing new interest and concern for health needs and inequalities of the whole global population, beyond the need to control specific issues such as epidemics and pandemics (Brown et al. 2006). Various health institutions and national and translational political bodies now use global health as a framing to discuss health policy and related issues. Similarly to planetary health, global health is based on a relatively recent realisation in the health sciences and epidemiology in particular—the idea that socio-economic factors have a crucial impact on the development of health and disease at the individual and population level, to the point that significant differences and inequalities in these socioeconomic factors can have a significant impact on health and disease. This might seem like a trivial realisation by epidemiologists and health policy-makers, considering that epidemiology is traditionally the area of biomedical research that directly focuses on the distribution and determinants of disease and health in populations (Broadbent 2013) and the historical development of the discipline has as such been tied to various public health measures and interventions on the environment, including socio-economic environments (Morabia 2004). Indeed, the study of socio-economic status is a historical feature of public health and arguably one of the defining characteristics at the origin of the field, 7 See Giroux (2021) on conceptual approaches and limits to the idea of populations being healthy.

116

S. CANALI

but the notions of social determinants of health is relatively recent and connected to research in the 1970s (such as the famous Whitehall studies on grade of employment and cardiovascular disease). However, similarly to planetary health, contemporary epidemiology and the focus on risk factors have rarely led to collaborations and interactions with the social sciences.The reason is evident when looking at data integration as the main focus of our analysis. Evidential standards tend to be significantly different between the health and social sciences, as contemporary epidemiological research is usually based on individuals and not populations as the main units of data collection and analysis (Kelly and Russo 2021), curation and annotation of metadata are differently advanced in the two disciplines (Boumans and Leonelli 2020), and the social sciences often lack the data and storing infrastructures that are typical of the biomedical sciences (Ankeny and Leonelli 2016). This has led to little similarity and compatibility at the level of data and evidence across the life, health, and social sciences. Epistemological and knowledge standards can vary greatly between the health and social sciences, with the latter focusing on qualitative and often descriptive goals while the former aim for the quantitative study of the determinants of health and disease and knowledge is normally presented in quantitative terms with the crucial goal of generalisation. As a result, the interactions between environments, societies, populations, and individuals are significantly understudied and rarely with the goal of causal inference in mind (Ghiara and Russo 2019).8 A general yet substantial disregard for socioeconomic determinants of health and disease is thus crucial background for the introduction of the concept of global health. But this is even more significant when considering a more general and widespread disregard for socio-economic features and causes of health and disease in other areas of the world than the West. The approach of global health is based on the use of health data from all the different populations of the globe for the study of population health. Global health is in this sense primarily a framework coming from health and economic policy, gaining its roots and background in planetary health, based on the economic and political role of the World Bank and the United Nations, and now emerging at the intersection of 8 The COVID-19 pandemic is a source of examples in this direction, where these differences are partly responsible for the fact that public health policy-making has rarely been grounded in social scientific knowledge and evidence (Lohse and Canali 2021).

WHICH INTEGRATION FOR HEALTH? COMPARING …

117

these and other public institutions such as the World Health Organisation and nongovernmental and translational players such as corporations and non-profit foundations (Reubi 2018). This is also the context where the collection and integration of global health data has taken place, with the development of large databases infrastructures and related data analysis, integration, and visualisation tools such as the Global Burden of Disease, a research centre and database located at the University of Washington and funded by the Gates Foundation. Yet this is also the context where we see the limitations of the approach to data integration of global health. The emergence and role of databases such as the Global Burden of Disease has been analysed by several historians and sociologists interested in the political and economic shifts around global health (Brown et al. 2006; Birn 2009; Reubi 2018).9 As discussed by Jean-Paul Gaudilliere and Camille Gasnier, the collection and use of global health data in the context of the Global Burden of Disease initiative was originally aimed at policy, particularly economic interventions for economic growth (Gaudilliere and Gasnier 2020). The initial development of the Global Burden of Disease was framed in particular around the comparison between different health interventions, with the goal of using health data as a basis to identify the most economically effective and efficient interventions. However, the shift towards the global health framework has also implied a shift in the goals of data collection and integration in the Global Burden of Disease, whereby global health data are increasingly used as single indicators of health and disease in different geographical distributions, rather than relating data to economic considerations about growth and efficiency (Gaudilliere and Gasnier 2020, 362–66). In other words, this has increasingly rendered global health data as primarily health data, to be used for the analysis and study of health and disease in the global population. A similar shift in the use of large data sets at the global level in epidemiology has taken place with the emergence of visualisation tools connected to these data, for instance with dedicated dashboards and maps that can track the spreading and impact of specific diseases and display these phenomena in constantly updated maps at a global scale.10

9 See https://www.thelancet.com/gbd (accessed October 2022). 10 These dashboards and maps have gained increasing political and epistemic importance

in the context of the COVID-19 pandemic, see a critical analysis of this by Susanne Bauer (2021).

118

S. CANALI

With global health, we thus see the repurposing of data collected at the global level for health purposes and their integration with other available data for the tracking and analysis of health and disease at the global scale. This shift has had significant consequences at the research and policy level: for instance, the use of data from the Global Burden of Disease has successfully shown the extent to which mental health and mental disorders are increasingly present in the Global South. Yet this repurposing for data integration has significant limitations too. The use of Global Burden of Disease data for the study of mental health has mostly not been based on local studies in the Global South, but rather on the “complex set of correlations between the burden of mental health disorders and various epidemiological, social and economical variables worked out in countries benefiting from more reliable statistics” (Gaudilliere and Gasnier 2020, 365). The type of data integration that has been elicited by global health initiatives is thus often based on data that are not actually global, which is a crucial problem for the intended aim of global health to study health and disease of the whole global population, including usually neglected populations. The problem here is that in many cases data on health and disease in areas of the world such as the Global South are just not available and initiatives to increase local data collection are scarce.11 More recently, a number of project in global health have tried to implement an approach presented as “precision global health” (Flahault et al. 2020; Sheath et al. 2020). The goal of these projects can be seen as precisely tapping into the issues discussed in this section—the lack of local and indigenous data on population health—through the collection of health data from digital devices such as smartphones, wearables, trackers. Setting aside ethical and social considerations on the use of these types of data as evidence for the study of population health, there is a broader issue that affects data integration for global health. In most cases, data from neglected areas of the world and low-resourced research environments tend to be perceived as low quality, which are not up to being integrated with other data (Leonelli 2017). This is among the reasons why more advanced technologies, such as digital devices, are proposed as ways of dealing with these issues. The use of digital technology as a way of including areas of the world that are underrepresented and excluded areas from health data collection is promising (Celi 2022), but the quality of data collected 11 See similar considerations by Rachel Ankeny on the need to bring data “out of the shadows” (Ankeny 2017).

WHICH INTEGRATION FOR HEALTH? COMPARING …

119

through these devices is often unclear and not transparent because of commercial interests. Data collection for global health has thus a crucial clash at its centre—the contrast between the need for more consideration of the global determinants of health and disease and evidential standards of what count as high-quality data across the globe.

6

Conclusions

In this chapter, I have analysed three approaches to the study of the relations between environment and population health, which have recently emerged in the context of various social and political discussions on health and biomedical research. Using the conceptual lens and methodological choice of focusing on data integration, I have looked at the exposome, planetary health, and global health as specific and distinct approaches to integrating different types of environmental and health data (see Table 1). With the introduction of these three concepts, we see new and different approaches to the integration of health and environmental data to study the relation between health and the environment, but also a conceptual and methodological expansion of what count as environment and health in relation to what count as environmental and health data and the ways these should be used and integrated for research in the health sciences. Table 1 Approaches to data integration analysed with respect to their specific features and limitations Approaches

Features

Limitations

Exposome

– Expansion of health data (e.g. molecular, omics, climatic) – Focus on genomics and omics techniques

Planetary health

– Expansion of environmental data (e.g. geographical, climatic, animal) – Focus on different environments and interactions between different species – Repurposing of health data at the global level (e.g. local data, data from the Global South) – Focus on health needs and inequalities of the global population

– Reductionism of genomic approaches – Genomic data as a proxy for environmental and social features – Gaps in data collection – Disciplinary boundaries – Missing data – Varying approaches towards data quality

Global health

120

S. CANALI

The exposome approach shows an expansion at the internal level of the environment and health, with the transfer of the genomic and molecular scaffolding for the study of internal exposure, their extension to epidemiology with an increasing focus on the individual level of environmental exposure, and thus a vertical expansion of the diversity of health data. As we have seen, this has significant consequences at the level of methodological approaches as well as disciplinary collaborations for the contemporary landscape of molecular and environmental epidemiology. The introduction of the concept of planetary health is a further expansion of these boundaries, but at the external level of the environment: a horizontal expansion of the types of environments that need to be analysed for health research, with a focus on the physical features of different environments and the interactions between different species that inhabit these environments and, therefore, also on different ways of collecting and integrating data on these new aspects of types of environment and health. As a consequence, the raise of this concept has significant consequences for the disciplinary boundaries of epidemiology, with the inclusion of further political and policy movements and the development of causal inference reflections on the influence of different environments on the health and disease of human and multispecies populations. The raise of the concept of global health has instead pushed for an expansion of the collection of health data—a horizontal expansion of the types of data that are considered necessary and the populations that need to be monitored and included in studies of population health and as such need to be integrated with environmental data. As a result, the extension of global health as a conceptual framework and approach is tied with the development of data infrastructures and partnerships across health and political institutions in different parts of the world. The focus on data integration sheds light on the concrete implications of following these conceptual and methodological frameworks, thus enabling more understanding of the epistemic implications of otherwise abstract and sometimes vague conceptualisations related to health, disease, and the environment. As we have seen throughout the chapter, this also includes shedding light on the limitations of these approaches. A common theme on limitations emerges as a result of the analysis in this chapter, which are related to the use of data as an asset to build collaborations between the health and environmental sciences. One of the consequences of shifts in the volume and diversity of scientific data in the last two decades has been the increasing value of scientific data

WHICH INTEGRATION FOR HEALTH? COMPARING …

121

as epistemic, social, and political assets (Leonelli 2019). In this sense it is unsurprising that changes in conceptual and methodological frameworks of the health sciences are crucially tied to choices of which data to integrate and how. Yet this renewed value of scientific data lies in the extraction of evidence and knowledge from data as an asset, which is not an automatic nor neutral act. As we have seen, the use of specific types of health and environmental data (e.g. planetary health and epidemiological data) is tied with methodological choices and epistemic assumptions of the collection and interpretation of the same data—data are no readymade solutions for the study of relations between environment and health and the collaboration between fields therein. The assumptions, methods, judgements that form the contextual features of e.g. molecular data need to be taken into account when data are integrated and aligned with other and existing assumptions, methods, judgements of the different disciplinary approaches in e.g. exposome research. This is even more significant when, as we have seen for instance with global data, data are not available and other data need to be repurposed for new uses. Data is hence a crucial asset for the study of the relations between environment and health in contemporary epidemiology—but attention to the contextual and epistemic implications of using different data is in turn crucial in order to fulfil these aims.

References Ankeny, Rachel A. 2017. ‘Bringing Data Out of the Shadows’. Science, Technology, & Human Values 42 (2): 306–10. https://doi.org/10.1177/016224 3916689138. Ankeny, Rachel A., and Sabina Leonelli. 2016. ‘Repertoires: A Post-Kuhnian Perspective on Scientific Change and Collaborative Research’. Studies in History and Philosophy of Science Part A 60 (December): 18–28. https:// doi.org/10.1016/j.shpsa.2016.08.003. Bauer, Susanne. 2013. ‘Modeling Population Health: Reflections on the Performativity of Epidemiological Techniques in the Age of Genomics’. Medical Anthropology Quarterly 27 (4): 510–30. https://doi.org/10.1111/maq. 12054. Bauer, Susanne. 2021. ‘Pandemic Infrastructure: Epidemiology as Compartmentalization’. Mefisto. Rivista Di Medicina, Filosofia, Storia 5 (1): 79–104. Birn, Anne-Emanuelle. 2009. ‘The Stages of International (Global) Health: Histories of Success or Successes of History?’ Global Public Health 4 (1): 50–68. https://doi.org/10.1080/17441690802017797.

122

S. CANALI

Boniolo, Giovanni, and Marco J. Nathan, eds. 2017. Philosophy of Molecular Medicine: Foundational Issues in Research and Practice. New York: Routledge, Taylor & Francis Group. Boumans, Marcel, and Sabina Leonelli. 2020. ‘From Dirty Data to Tidy Facts: Clustering Practices in Plant Phenomics and Business Cycle Analysis’. In Data Journeys in the Sciences, edited by Sabina Leonelli and Niccolò Tempini, 79– 101. Cham: Springer International Publishing. https://doi.org/10.1007/ 978-3-030-37177-7_5. Brigandt, Ingo. 2010. ‘Beyond Reduction and Pluralism: Toward an Epistemology of Explanatory Integration in Biology’. Erkenntnis 73 (3): 295– 311. https://doi.org/10.1007/s10670-010-9233-3. Broadbent, Alex. 2013. Philosophy of Epidemiology. London: Palgrave Macmillan UK. https://doi.org/10.1057/9781137315601. Brown, Theodore M., Marcos Cueto, and Elizabeth Fee. 2006. ‘The World Health Organization and the Transition From “International” to “Global” Public Health’. American Journal of Public Health 96 (1): 62–72. https:// doi.org/10.2105/AJPH.2004.050831. Canali, Stefano. 2019. ‘Evaluating Evidential Pluralism in Epidemiology: Mechanistic Evidence in Exposome Research’. History and Philosophy of the Life Sciences 41 (1): 4. https://doi.org/10.1007/s40656-019-0241-6. Canali, Stefano. 2020a. ‘What Is New about the Exposome? Exploring Scientific Change in Contemporary Epidemiology’. International Journal of Environmental Research and Public Health 17 (8): 2879. https://doi.org/10.3390/ ijerph17082879. Canali, Stefano. 2020b. ‘Making Evidential Claims in Epidemiology: Three Strategies for the Study of the Exposome’. Studies in History and Philosophy of Science Part C: Studies in History and Philosophy of Biological and Biomedical Sciences 82 (August): 101248. https://doi.org/10.1016/j.shpsc.2019. 101248. Canali, Stefano, and Sabina Leonelli. 2022. ‘Reframing the Environment in DataIntensive Health Sciences’. Studies in History and Philosophy of Science 93 (June): 203–14. https://doi.org/10.1016/j.shpsa.2022.04.006. Celi, Leo Anthony. 2022. ‘PLOS Digital Health, a New Journal Driving Transformation in the Delivery of Equitable and Unbiased Healthcare’. PLOS Digital Health 1 (1): e0000009. https://doi.org/10.1371/journal.pdig.000 0009. Chadeau-Hyam, Marc, Toby J. Athersuch, Hector C. Keun, Maria De Iorio, Timothy M.D. Ebbels, Mazda Jenab, Carlotta Sacerdote, Stephen J Bruce, Elaine Holmes, and Paolo Vineis. 2011. ‘Meeting-in-the-Middle Using Metabolic Profiling – a Strategy for the Identification of Intermediate Biomarkers in Cohort Studies’. Biomarkers 16 (1): 83–88. https://doi.org/ 10.3109/1354750X.2010.533285.

WHICH INTEGRATION FOR HEALTH? COMPARING …

123

Chadeau-Hyam, Marc, Gianluca Campanella, Thibaut Jombart, Leonardo Bottolo, Lutzen Portengen, Paolo Vineis, Benoit Liquet, and Roel C.H. Vermeulen. 2013. ‘Deciphering the Complex: Methodological Overview of Statistical Models to Derive OMICS-Based Biomarkers: Statistical Approaches for OMICS-Based Biomarkers’. Environmental and Molecular Mutagenesis 54 (7): 542–57. https://doi.org/10.1002/em.21797. Dupré, John. 1996. The Disorder of Things: Metaphysical Foundations of the Disunity of Science. Cambridge, Mass.: Harvard Univ. Press. Flahault, Antoine, Jürg Utzinger, Isabella Eckerle, Danny J Sheath, Rafael Ruiz de Castañeda, Isabelle Bolon, Nefti-Eboni Bempong, and Fred Andayi. 2020. ‘Precision Global Health for Real-Time Action’. The Lancet Digital Health 2 (2): e58–59. https://doi.org/10.1016/S2589-7500(19)30240-7. Fleming, Lora, Niccolò Tempini, Harriet Gordon-Brown, Gordon L. Nichols, Christophe Sarran, Paolo Vineis, Giovanni Leonardi, et al. 2017. ‘Big Data in Environment and Human Health’. In Oxford Research Encyclopedia of Environmental Science, by Lora Fleming, Niccolò Tempini, Harriet GordonBrown, Gordon L. Nichols, Christophe Sarran, Paolo Vineis, Giovanni Leonardi, et al. Oxford University Press. https://doi.org/10.1093/acrefore/ 9780199389414.013.541. Gaudilliere, Jean-Paul, and Camille Gasnier. 2020. ‘From Washington DC to Washington State: The Global Burden of Diseases Data Basis and the Political Economy of Global Health’. In Data Journeys in the Sciences, edited by Sabina Leonelli and Niccolò Tempini, 351–69. Cham: Springer International Publishing. https://doi.org/10.1007/978-3-030-37177-7_18. Ghiara, Virginia, and Federica Russo. 2019. ‘Reconstructing the Mixed Mechanisms of Health: The Role of Bio- and Sociomarkers’. Longitudinal and Life Course Studies 10 (1): 7–25. https://doi.org/10.1332/175795919X15468 755933353. Gibbon, Sahra, Barbara Prainsack, Stephen Hilgartner, and Janelle Lamoreaux. 2020. Routledge Handbook of Genomics, Health and Society. London: Routledge Giroux, Élodie. 2021. ‘Can Populations Be Healthy? Perspectives from Georges Canguilhem and Geoffrey Rose’. History and Philosophy of the Life Sciences 43 (4): 111. https://doi.org/10.1007/s40656-021-00463-x. Giroux, Élodie, Yohan Fayet, and Thibaut Serviant-Fine. 2021. ‘L’Exposome: Tensions entre holisme et réductionnisme’. médecine/sciences 37 (8–9): 774– 78. https://doi.org/10.1051/medsci/2021092. Golub, Todd. 2010. ‘Counterpoint: Data First’. Nature 464 (7289): 679– 679. https://doi.org/10.1038/464679a. Green, Sara, and Henrik Vogt. 2016. ‘Personalizing Medicine: Disease Prevention in Silico and in Socio’. HUMANA.MENTE Journal of Philosophical Studies 9 (30): 42.

124

S. CANALI

Gulliver, John, David Morley, Chrissi Dunster, Adrienne McCrea, Erik van Nunen, Ming-Yi Tsai, Nicoltae Probst-Hensch, et al. 2018. ‘Land Use Regression Models for the Oxidative Potential of Fine Particles (PM 2.5) in Five European Areas’. Environmental Research 160 (January): 247–55. https:// doi.org/10.1016/j.envres.2017.10.002. Hilgartner, Stephen. 2017. Reordering Life: Knowledge and Control in the Genomics Revolution. Inside Technology. Cambridge, Massachusetts: The MIT Press. Hogle, Linda F. 2016. ‘Data-Intensive Resourcing in Healthcare’. BioSocieties 11 (3): 372–93. https://doi.org/10.1057/s41292-016-0004-5. Holmberg, Christine, Christine Bischof, and Susanne Bauer. 2013. ‘Making Predictions: Computing Populations’. Science, Technology, & Human Values 38 (3): 398–420. https://doi.org/10.1177/0162243912439610. Horton, Richard, Robert Beaglehole, Ruth Bonita, John Raeburn, Martin McKee, and Stig Wall. 2014. ‘From Public to Planetary Health: A Manifesto’. The Lancet 383 (9920): 847. https://doi.org/10.1016/S0140-673 6(14)60409-8. Illari, Phyllis, and Federica Russo. 2016. ‘Information Channels and Biomarkers of Disease’. Topoi 35 (1): 175–90. https://doi.org/10.1007/s11245-0139228-1. Kelly, Michael P, and Federica Russo. 2021. ‘The Epistemic Values at the Basis of Epidemiology and Public Health’. Mefisto. Rivista Di Medicina, Filosofia, Storia 5 (1): 105–20. Klingwort, Jonas, and Rainer Schnell. 2020. ‘Critical Limitations of Digital Epidemiology’. Survey Research Methods, June, 95–101. https://doi.org/10. 18148/SRM/2020.V14I2.7726. Landecker, Hannah, and Aaron Panofsky. 2013. ‘From Social Structure to Gene Regulation, and Back: A Critical Introduction to Environmental Epigenetics for Sociology’. Annual Review of Sociology 39 (1): 333–57. https://doi.org/ 10.1146/annurev-soc-071312-145707. Leonelli, Sabina. 2013. ‘Integrating Data to Acquire New Knowledge: Three Modes of Integration in Plant Science’. Studies in History and Philosophy of Science Part C: Studies in History and Philosophy of Biological and Biomedical Sciences 44 (4): 503–14. https://doi.org/10.1016/j.shpsc.2013.03.020. Leonelli, Sabina. 2016. Data-Centric Biology: A Philosophical Study. Chicago London: The University of Chicago Press. Leonelli, Sabina. 2017. ‘Global Data Quality Assessment and the Situated Nature of “Best” Research Practices in Biology’. Data Science Journal 16 (June): 32. https://doi.org/10.5334/dsj-2017-032. Leonelli, Sabina. 2019. ‘Data — from Objects to Assets’. Nature 574 (7778): 317–20. https://doi.org/10.1038/d41586-019-03062-w.

WHICH INTEGRATION FOR HEALTH? COMPARING …

125

Leonelli, Sabina, and Niccolò Tempini, eds. 2020. Data Journeys in the Sciences. Cham: Springer International Publishing. https://doi.org/10.1007/978-3030-37177-7. Lohse, Simon, and Stefano Canali. 2021. ‘Follow *the* Science? On the Marginal Role of the Social Sciences in the COVID-19 Pandemic’. European Journal for Philosophy of Science 11 (4): 99. https://doi.org/10.1007/ s13194-021-00416-y. Mitchell, Sandra D., and Angela M. Gronenborn. 2017. ‘After Fifty Years, Why Are Protein X-Ray Crystallographers Still in Business?’ The British Journal for the Philosophy of Science 68 (3): 703–23. https://doi.org/10.1093/bjps/ axv051. Mitchell, Sandra D., and Michael R. Dietrich. 2006. ‘Integration without Unification: An Argument for Pluralism in the Biological Sciences’. The American Naturalist 168 (S6): S73–79. https://doi.org/10.1086/509050. Mittelstadt, Brent, Justus Benzler, Lukas Engelmann, Barbara Prainsack, and Effy Vayena. 2018. ‘Is There a Duty to Participate in Digital Epidemiology?’ Life Sciences, Society and Policy 14 (1): 9. https://doi.org/10.1186/s40504-0180074-1. Morabia, Alfredo, ed. 2004. A History of Epidemiologic Methods and Concepts. Basel: Birkhäuser Basel. https://doi.org/10.1007/978-3-0348-7603-2. Müller-Wille, Staffan, and Isabelle Charmantier. 2012. ‘Natural History and Information Overload: The Case of Linnaeus’. Studies in History and Philosophy of Science Part C: Studies in History and Philosophy of Biological and Biomedical Sciences 43 (1): 4–15. https://doi.org/10.1016/j.shpsc.2011. 10.021. Myers, Samuel S, Jeremy I Pivor, and Antonio M Saraiva. 2021. ‘The São Paulo Declaration on Planetary Health’. The Lancet, October, S0140673621021814. https://doi.org/10.1016/S0140-6736(21)02181-4. Parker, Wendy. 2018. ‘Climate Science’. In Stanford Encyclopedia of Philosophy. Vol. Summer 2018 Edition. Stanford (CA): Metaphysics Research Lab, Stanford University. https://plato.stanford.edu/archives/sum2018/entries/ climate-science/. Plutynski, Anya. 2022. ‘Why Precision Oncology Is Not Very Precise (and Why This Should Not Surprise Us)’. In Personalized Medicine in the Making. Philosophical Perspectives from Biology to Healthcare, edited by Chiara Beneduce and Marta Bertolaso, Springer, Cham. 3–21. Prainsack, Barbara. 2020. ‘The Meaning and Enactment of Openness in Personalised and Precision Medicine’. Science and Public Policy 47 (5): 647–54. https://doi.org/10.1093/scipol/scaa013. Rappaport, Stephen M., and Martyn T. Smith. 2010. ‘Environment and Disease Risks’. Science 330 (6003): 460–61. https://doi.org/10.1126/science.119 2603.

126

S. CANALI

Reubi, David. 2018. ‘Epidemiological Accountability: Philanthropists, Global Health and the Audit of Saving Lives’. Economy and Society 47 (1): 83–110. https://doi.org/10.1080/03085147.2018.1433359. Richardson, Sarah S., and Hallam Stevens, eds. 2015. Postgenomics: Perspectives on Biology after the Genome. Durham: Duke University Press. Russo, Federica. 2009. ‘Variational Causal Claims in Epidemiology’. Perspectives in Biology and Medicine 52 (4): 540–54. https://doi.org/10.1353/pbm.0. 0118. Russo, Federica, and Paolo Vineis. 2016. ‘Opportunities and Challenges of Molecular Epidemiology’. In Philosophy of Molecular Medicine, edited by Giovanni Boniolo and Marco J. Nathan. New York: Routledge. Russo, Federica, and Jon Williamson. 2007. ‘Interpreting Causality in the Health Sciences’. International Studies in the Philosophy of Science 21 (2): 157–70. https://doi.org/10.1080/02698590701498084. Salathé, Marcel. 2018. ‘Digital Epidemiology: What Is It, and Where Is It Going?’ Life Sciences, Society and Policy 14 (1): 1. https://doi.org/10.1186/ s40504-017-0065-7. Sheath, Danny J., Rafael Ruiz de Castañeda, Nefti-Eboni Bempong, Mario Raviglione, Catherine Machalaba, Michael S. Pepper, Effy Vayena, et al. 2020. ‘Precision Global Health: A Roadmap for Augmented Action’. Journal of Public Health and Emergency 4 (March): 5–5. https://doi.org/10.21037/ jphe.2020.01.01. Shostak, Sara, and Margot Moinester. 2015. ‘The Missing Piece of the Puzzle? Measuring the Environment in the Postgenomic Moment’. In Postgenomics: Perspectives on Biology after the Genome, edited by Sarah S. Richardson and Hallam Stevens. London: Duke University Press. Stegenga, Jacob, Ashley Graham, Serife ¸ Tekin, Saana Jukola, and Robin Bluhm. 2017. ‘New Directions in Philosophy of Medicine’. In The Bloomsbury Companion to Contemporary Philosophy of Medicine, edited by James Marcum. Bloomsbury Academic. Strasser, Bruno J. 2019. Collecting Experiments: Making Big Data Biology. Chicago: The University of Chicago Press. Valles, Sean A. 2019. Philosophy of Population Health Science: Philosophy for a New Public Health Era. Routledge. Valles, Sean A. 2021. ‘A Pluralistic and Socially Responsible Philosophy of Epidemiology Field Should Actively Engage with Social Determinants of Health and Health Disparities’. Synthese 198: 2589–2611. https://doi.org/ 10.1007/s11229-019-02161-5. Vogt, Henrik, Bjørn Hofmann, and Linn Getz. 2016. ‘The New Holism: P4 Systems Medicine and the Medicalization of Health and Life Itself’. Medicine, Health Care and Philosophy 19 (2): 307–23. https://doi.org/10.1007/s11 019-016-9683-8.

WHICH INTEGRATION FOR HEALTH? COMPARING …

127

Warde, Paul, Libby Robin, and Sverker Sörlin. 2018. The Environment: A History of the Idea. Baltimore, Maryland: John Hopkins University Press. Weinberg, Robert. 2010. ‘Point: Hypotheses First’. Nature 464 (7289): 678– 678. https://doi.org/10.1038/464678a. Wild, Christopher Paul. 2012. ‘The Exposome: From Concept to Utility’. International Journal of Epidemiology 41 (1): 24–32. https://doi.org/10.1093/ ije/dyr236. Wild, Cristopher Paul. 2005. ‘Complementing the Genome with an “Exposome”: The Outstanding Challenge of Environmental Exposure Measurement in Molecular Epidemiology’. Cancer Epidemiology Biomarkers & Prevention 14 (8): 1847–50. https://doi.org/10.1158/1055-9965.EPI-05-0456. Worrall, John. 2002. ‘What Evidence in Evidence-Based Medicine?’ Philosophy of Science 69 (S3): S316–30. https://doi.org/10.1086/341855. Worrall, John. 2007. ‘Evidence in Medicine and Evidence-Based Medicine’. Philosophy Compass, 2 (6) 981–1022.

A Critical Assessment of Exposures Integration in Exposome Research

Élodie Giroux

1

Introduction1

The exposome, defined by Christopher Wild (2005, 1848) as that which “encompasses life-course environmental exposures (including lifestyle factors), from the period onwards”, aims to establish itself in environmental health research as a new federative approach. In a post-genomic context, and in a similar way to “precision medicine”,2 the exposome benefits from a form of economy of promises (Joly 2010) and a “pull effect”, with many public and private investments. The approach for which this concept is the vehicle is seductive in many respects for 1 This chapter is a translation with some modifications and additions of an article published in French in Lato Sensu en 2021 (see Giroux 2021a). 2 Precision medicine is an extension of so-called personalized medicine, which aims to adapt prevention and treatment to the individual characteristics of patients. To this end, precision medicine draws on the collection and analysis of massive individual data (National Research Council 2011; Collins and Varmus 2015).

É. Giroux (B) Institut de Recherches Philosophiques de Lyon, Université Jean Moulin Lyon 3, Lyon, France e-mail: [email protected]

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 É. Giroux et al. (eds.), Integrative Approaches in Environmental Health and Exposome Research, https://doi.org/10.1007/978-3-031-28432-8_6

129

130

É. GIROUX

anyone who defends the importance of adopting a comprehensive and multifactorial view of disease causation. Historically, taking the environment into account is associated with a conception of health and disease that is considered holistic and with preventive actions deriving more from public health than from clinical medicine. The “science of the exposome” (Vineis 2018), a comprehensive and integrative vision of exposures, promises a better explanation of chronic diseases, the objective of which is to improve prevention in public health (Wild 2005, 2012; Juarez et al. 2014) as well as in the context of personalized or precision medicine (Rappaport and Smith 2010; Rappaport 2011). At first sight, then, the exposome—presented as “complementing the genome” (Wild 2005)—seems to overcome the limits of an approach that is restricted to the biological, and even the genetic or molecular, dimensions of health and disease aetiology. The affirmation of the integrative, systematic, comprehensive and interdisciplinary scope of this emergent field is very pronounced amongst all its advocates.3 But the examination of the diverse approaches and studies associated with the notion of exposome reveals that the meanings given to the notions of “holism” and “integration” are far from being univocal. Definitions of the exposome are themselves varied and leave scope for a highly plastic concept, lending itself to different interpretations by the different disciplines in environmental health. The notion of “integration” covers different meanings and diverse degrees. Beyond a purely rhetorical usage, it is not easy to disentangle what exactly the exposome aims to integrate (exposure data or variables? exposure and/or effects on health? methods? explanations? disciplines? the biomedical frameworks? etc.) and to uncover the type (explanative? reductive? etc.) and intensity (strong or weak) of the envisaged integration. An analysis of the publications on the exposome (2005–2021) actually reveals a tension between two interpretative poles, one more reductionist and the other more holistic, which may be considered as two research orientations of the field (Giroux et al. 2021). On the one hand, the exposome appears as a prolongation of genomic or precision medicine. The central objective is, by analogy with sequencing the genome, to make the measurements of exposures more precise thanks to the identification 3 See for example the editorial by Gary Miller justifying the creation of a dedicated journal “Exposome” at Oxford University Press: “Exposome: a new field, a new journal”, Exposome, 2021, vol. 1, N° 1.

A CRITICAL ASSESSMENT OF EXPOSURES INTEGRATION …

131

of biomarkers at the level of the internal biochemical environment of the body; this seems to lead to a form of “internalization” of the environment and to its “molecularization”. And on the other hand, the exposome is envisaged as the opportunity to adopt a multidimensional perspective on exposures and thus to bring together different disciplines that had hitherto cooperated only little: molecular epidemiology and toxicology, but also social epidemiology, geography, environmental sociology and justice, etc. Some go as far as to defend the idea that the exposome is the occasion to develop “a holistic model of environmental health” (Senier et al. 2017, 2). The exposome is also thought to provide the opportunity to further integrate social and biological exposures and to better explain the interaction of these various exposures in the pathological effect (Vineis 2018, 360). As is often the case in the sciences, the ambiguity of the exposome concept and of the integrative scope associated with it further nourishes its promises. And this ambiguity is no doubt irreducible at this early stage, at which it is difficult, through lack of distance, to say what the fecundity of this research will be and what effective applications might arise. The objective of this article cannot therefore be to predict whether the most holistic promises will be fulfilled or to determine whether they are utopian or purely instrumental to the benefit of reductionist genomic and biological logic. Neither is the objective to put forward a clear and univocal definition of the exposome concept. A certain plasticity of the concept is no doubt necessary for the development of the field.4 More modestly, this contribution aims, on the basis of an examination of the literature, of an analysis of the first generation of exposomic studies (2010–20), and of interviews carried out with epidemiologists engaged in the research, to consider the plurality of orientations and research agendas and to make more explicit their presuppositions. In this way, a review will be carried out of the promises and realizations made with regard to the

4 Michel Morange clearly showed with respect to the concept of the gene that its fuzzy and elusive character is not in itself a problem. It remains possible to put forward more precise local and contextual definitions. Above all, he underlined that the plasticity of scientific concepts is in reality necessary for scientific work. “To give precise definitions of scientific concepts would be to rigidify them, to prevent this permanent reorganization of knowledge which is the motor of scientific progress. Because a concept is fuzzy, it is rich in explanatory potential” (Morange 1998, 39, our translation). I thank Pierre-Olivier Méthot for having provided this reference.

132

É. GIROUX

integration of exposures, while also bringing a critical and interrogative perspective to the discussion. This will also be the occasion for a more general philosophical contribution to the reflection on integration in philosophy of science, its nature and its relevance, between pluralism and unification (see for example: Plutynski 2013; Longino 2013; Brigandt 2010; Mitchell 2002). Focusing on the epistemological dimension of the specific issue of integration and of the tension between holistic and reductionist approaches in this emergent field, we will not engage here with the ethics of environmental health science and exposomics, even if some of our reflections could be relevant to it. The analysis begins by extracting a common core of the current use of the “exposome” concept, despite the diversity of definitions. Then, after having contextualized the concept’s emergence, what is truly innovative amongst the many claims made about it is identified. Starting from an examination of some of the most important of the first generation of exposome studies, an analysis is carried out of two main orientations. On the one hand, one prioritize greater precision in the measurements made at the biological level, and on the other, the integration of social factors. It is with this question of the project of integrating biological and social exposures that the third and longest section of this contribution is concerned. In this case, the type of integration appears to be explanative and strong in its intensity. The objective is to consider whether and in what manner this orientation of the exposome research renews the ways that epidemiology has sought to integrate the biological and the social in the study of disease aetiology. The relevance of using the “biomarker approach” and the mechanistic view of causation in the context of a strong explanative integrative perspective is then questioned. Finally, the question of whether the goal of a strong integration is the most relevant for developing preventive action in the domain of environmental health is shortly addressed.

2 2.1

Context and Promises of the Exposome Limits of the Traditional Approaches to Environmental Exposures

Chronic diseases have become dominant, through their contribution to morbidity and mortality. In a context in which hopes of a better characterisation of their aetiology through genetic knowledge have met with

A CRITICAL ASSESSMENT OF EXPOSURES INTEGRATION …

133

disappointment, health sciences are according greater importance to the environment with a view to refining our understanding of the origin and the development of these diseases. One of the principal promoters of the exposome, Stephen Rappaport, underscores the fact that non-genetic (i.e. to him, environmental) factors explain 90% of the risks of chronic disease (2011, 6). Many advocates of the exposome concept make this claim, even if the precise quantification of the relevant percentage remains a subject of debate (Saracci and Vineis 2007).5 A consensus has nevertheless been established that the genetic contribution to aetiology is relatively small in the case of most chronic diseases and that refining their understanding calls for the study of interactions between genes and environment. But the measurement of environmental factors and the evaluation of their impact on health are particularly complex and difficult to carry out. Epidemiology and toxicology are the two main disciplines that have, throughout the twentieth century, made it possible to study and measure the health impact of some environmental exposures. But they are limited by their respective methodologies, especially when it is a question of studying several exposures at low doses and their long-term toxicity. Yet those types of exposure and their mixture are in play in a great number of the chronic diseases prevalent today. The development of causal epidemiology, since the middle of the twentieth century, has made it possible to identify the major risk factors for cardiovascular diseases and cancers, that is to say, the factors whose “relative risk” (i.e. a measure of the strength of the association) is high. Nevertheless, these factors are essentially behavioural and physiological, measured at the individual level, and then different from environmental factors in the strict sense of external factors distinct from the individual organism or the host6 (Giroux 2013). The precise measurement of the

5 The evaluation of the proportion of diseases due to the environment is particularly complex, and each researcher puts forward their own numbers. See Slama (2017). Moreover, a radical criticism has been made by the social epidemiologist Nancy Krieger (2017)—an author who will be central in section 4 for our critical analysis of the biosocial integration in exposomics—in a paper arguing that environmental causes and other health causes are not additive in a way that makes it possible to attribute X% to genetics and X% to the environment (many thanks to the anonymous reviewer for this reference). 6 The question of the inclusion of behavioural factors in the category of environmental factors is a subject of debate. Lifestyle is sometimes included in the notion of environment and sometimes distinguished as an entirely separate category (as in the WHO categorization).

134

É. GIROUX

exposure of individuals to environmental factors, like air pollution, is in fact more complex, and the evaluation of their impact on the health of individuals is often indirect. Indeed, causal epidemiology is based on an observational approach and comparative studies of well-defined populations; but causal inference based on the identification of statistical associations poses profound difficulties. The problem of bias is of particular concern, and poses an even greater danger when the statistical associations are weak. Very large sample groups are required. As a consequence, epidemiologists have taken an interest in certain professional exposures at high doses, in a context in which it is possible to draw on large cohorts. But an environmental factor, taken in isolation, and outside of certain accident situations (chemical factories, for example) or in the case of specific diseases (e.g. cancer of the pleura from asbestos, saturnism from lead), only plays a partial role in the aetiology of multifactorial diseases. And it is often a question of low doses and long-term toxicities. For example, in the case of endocrine disruptors, an exposure in utero may be responsible for effects in adulthood (Kortenkamp et al. 2011). And risks exhibiting little specificity are much harder to apprehend, but are nevertheless important for public health, when a large number of individuals are exposed. But above all, the type of statistical evidence brought by comparative studies of “risk factor epidemiology”— also labelled “black box epidemiology” (Vandenbroucke 1988)—by its critics, is considered insufficient for causal inference. Philosophers, trying to explain the conception of researchers in medicine and epidemiology, have argued that the identification of mechanisms linking exposure to its effect is necessary for a causal inference to be possible (Russo and Williamson 2007). Causal inference in medicine would thus be based on an evidential pluralism, such that one must have evidence of both physical mechanism and probabilistic dependencies (or difference-making evidence). As for toxicology, or the science of poisons, it uses the experimental method and animal models to analyse the interaction between a toxic agent and a target. It studies the health effect of environmental contaminants and makes it possible to dissect, stage by stage, and organ by organ, the mechanisms involved in this effect at different scales, from the molecule to the organism (Slama 2017, 77). It allows one to identify certain mechanisms explaining the statistical association observed in epidemiology. But its results are difficult to extrapolate to humans under real exposure conditions. Moreover, the dominant idea which has for a

A CRITICAL ASSESSMENT OF EXPOSURES INTEGRATION …

135

long time structured toxicology that “the dose makes the poison” has recently been challenged. The environmental exposures at issue today in the health of populations are often low but can have a significant health impact over the long term. Further, toxicology, like epidemiology, limits itself to an approach that studies one exposure at a time and targeting a restricted set of substances known for being biologically active (Rappaport 2011), which does not allow one to take account of long-term cumulative and dynamic effects. But that leads one to overlook numerous potentially noxious substances as well as the so-called cocktail effect. Thus, toxicology and epidemiology face important epistemological issues when identifying and measuring health exposures and their health effects.7 2.2

Diverse Definitions of the Exposome

Before clarifying how the concept of the exposome aims to address the difficulties mentioned above, let us first consider its definition, which is far from being univocal. The term “exposome” was introduced in 2005 by the cancer and molecular epidemiologist, Christopher Wild, in an editorial of the journal Cancer Epidemiology Biomarkers and Prevention (of which he is the co-editor in chief) entitled: “Complementing the Genome with an ‘exposome’: the outstanding challenge of environmental exposure measurement in molecular epidemiology”. The exposome is defined here as that which “encompasses life-course environmental exposures (including lifestyle factors), from the prenatal period onwards”. Great importance is accorded to the dynamics of exposures whose effects may be cumulative over the long term and whose impact may increase during the critical windows of the foetal stage and of development. Moreover, the objective is to develop tools for measuring exposures that would get as close as possible to the precision obtained in the characterisation of the genome. It is important to note right at the outset a tension between a totalizing comprehensive aim (“complement”) and a methodology aiming at a form of precision analogous to that obtained in genome sequencing (Giroux 2021b). Then, in 2010, the exposome notion was taken up and refined in the journal Science by the chemist, Stephen Rappaport and the chemist 7 On the more general question of the “scientific” vulnerability of the environmental sciences, which make them particularly “exposed” to conflicts of political and economic value, see Shostak (2013).

136

É. GIROUX

and toxicologist, Martyn Smith, who centre the definition on internal exposures, i.e. the chemical components that are biologically active inside the body (Rappaport and Smith 2010). To them exposures are “not restricted to chemicals (toxicants) entering the body from air, water, or food, for example, but also include chemicals produced by inflammation, oxidative stress, lipid peroxidation, infections, gut flora, and other natural processes”. In quite a counter-intuitive way with respect to current usage in epidemiology of the concepts of environment and exposure, conceived as external to the body of the individual, the most “relevant” environment is explicitly considered here as “the body’s internal chemical environment” (Rappaport 2011, 6). This concept of the exposome closely resembles Claude Bernard’s physiological concept of the “internal milieu”. In 2012, Wild published an article, “The exposome: from concept to utility” (2012), addressing a readership of non-specialist epidemiologists (International Journal of Epidemiology), in which he put forward a categorization integrating the notion of “internal environment” but which maintains the importance accorded to external exposures, themselves divided between “general” external and “specific” external. These three principal domains of the exposome constitute the triad to which numerous publications on the exposome refer, and which is reminiscent of the traditional epidemiological triad of “agent-host-environment”: – Internal (host ): “metabolism, endogenous hormones, body morphology, physical activity, gut microflora, inflammation, lipid peroxidation, oxidative stress, ageing, etc.” – Specific external (agent ): “radiation, infectious agents, chemical contaminants and environmental pollutants, diet, lifestyle factors (e.g. tobacco, alcohol), occupation, medical interventions, etc.” – General external (environment ): “social capital, education, financial status, psychological and mental stress, urban–rural environment, climate, etc.” If this division aims to go beyond former categorizations which distinguish the physical, the natural and the social environment, without including the internal environment of the organism, neither its pertinence nor its ground is the subject of sustained discussion in the literature. Beside this, the notion of “environment” tends to disappear in the

A CRITICAL ASSESSMENT OF EXPOSURES INTEGRATION …

137

language of the promoters of the exposome, replaced by that of “exposure”. In particular, the notion of environment is judged too restrictive because too closely associated in people’s minds to that of “pollution” (Rappaport et al. 2014, 769). In 2014, Gary Miller, toxicologist, pharmacologist and author of the first book on the exposome (Miller 2014), and Dean Jones, a biochemist, put forward an enlarged definition of the internal exposome in another less cited articled: “The nature of nurture: refining the definition of exposome” (Miller and Jones 2014), published in Toxicological Sciences. This definition is based on three additions with respect to the definition of Wild: – the cumulative biological responses – the behaviour, in a broadest sense beyond lifestyle – endogenous processes, whose advantage is to provide the evidence of an actual effect even decades after exposure The idea is to take account of the fact that the effect of the exposure can be considerably modified by the endogenic processes of the body. Thus, what matters is not so much the external exposure as the way that the individual organism reacts to it. These authors leave room for epigenetic alterations in the internal exposome, considering that they are as important as chemical alterations. The exposome is presented as “an integrated science of nurture”, making it possible to give “a biological index” which would complete the science of the innate. They then redefine the exposome as “the cumulative measure of environmental influence and associated biological responses throughout the lifespan, including exposures from the environment, diet, behavior, and endogenous processes” (2014, 2). The internal exposome again plays a dominant role here and the external environment is only one exposure amongst others. Thus, the definitions seem to vary as a function of the disciplinary origin of the definer and/or the targeted readership: the epidemiologist, Wild, is concerned with a wide and comprehensive integrative sense of the exposome, covering the set of determinants and factors to which the general epidemiologist is habitually interested, whereas the chemists and toxicologists, Rappaport, Smith and Miller, centre their definition on the

138

É. GIROUX

internal exposome, which is quite natural, as long as they stay anchored in the logic of their own disciplines. 2.3

A Common Core

Despite the diversity present in definitions of the exposome, two principal dimensions are always associated with it and can be considered to constitute the “signature” of the concept. First, it is a question of studying several exposures at the same time and breaking with the approach of targeting one exposure at a time. Indeed, “most notably, exposome research embraces all exposures (or at least as many as possible) rather than focusing on a particular exposure or class of environmental chemicals” (Buck et al. 2017). Second, it is a question of studying exposures by taking account of their cumulative character over time. The importance of longitudinal cohort studies is underlined in order to grasp the cumulative nature, throughout the life course, of different environmental factors, as well as the latency period preceding the emergence of numerous chronic diseases. Such an approach makes possible discoveries about the late health impact of exposures occurring early in life. These two dimensions, which may be respectively characterized as synchronic and diachronic (or biographical), aim to resolve some of the difficulties and limits of the study of the environmental exposures mentioned above. To these two dimensions one may add the improvement of the characterisation and the understanding of pathological mechanisms, promised by the identification of biomarkers and the use of “omic” technologies. Biomarkers, whether they are of exposure, of effect of exposure (indication of a specific disease for example) or of susceptibility, are considered to guarantee a more precise and objective account of exposures and to offer a better understanding of the causal pathway linking the exposure to the disease. By tracking down the exposure at the molecular level, one may better understand how it influences the development of a disease (Vineis et al. 2017b; Vineis 2018; Rappaport 2018; Illari and Russo 2016). The “meet-in-the-middle” approach is particularly solicited (Wild 2012; Wild et al. 2013). It allows reinforcing the biological plausibility of the association between an exposure and a disease by identifying intermediate biomarkers between the exposure and the disease. This approach is thought to make it possible to open the “black box” of “risk factor epidemiology”.

A CRITICAL ASSESSMENT OF EXPOSURES INTEGRATION …

139

Thus we have described three main dimensions of the exposome as constituting its common core, despite the wide diversity of attempts to define it. It is then possible to ask whether and in what extent those dimensions are new for epidemiology? 2.4

What is Really New

The place accorded to the biomarker approach and the hopes of overcoming the limits of extrapolating from animals models to humans raised by a greater precision of measurements at the molecular level were already at the centre of molecular epidemiology from the 1980s onwards (Schulte and Perera 1998). This sub-discipline of epidemiology had thus already embraced the objective of mitigating certain of the limits mentioned above and bringing epidemiology closer to the experimental methods of the laboratory (Canali 2019; Shostak 2010). The notion of the exposome was introduced by a molecular epidemiologist (Wild) and belongs to this trend. Similarly, the diachronic dimension, which seeks to take account of exposures over the course of a lifetime, had made its way into epidemiology in the 1990s via the development of “life course epidemiology” (Ben-Schlomo 1997), which draws on the concept of “life course” developed in sociology. In the 1980s, researchers had revealed that birth weight in the intra-uterine period has an influence on cardiovascular risk in adulthood (Barker et al. 1989). This research led to the concept of the “Developmental Origins of Health and Disease” or DOHaD. Following on from this, what is new in the concept of the exposome would appear primarily to concern the synchronic dimension: insistence on integration, understood here as the simultaneous taking into account of multiple exposures, with exposures being understood in a very broad sense. The question of interdisciplinarity and the role of biostatistical and bioinformatic methods in the context of the use of “omic” technologies to measure, characterize and model multiple exposures are considered to play a determining role (Wild 2005, 2012). The focus on the internal exposome appears central in this context. There also arises the challenge of integrating data whose source (for example: questionnaires, blood samples, etc.) and level (molecular, organic, populational, etc.) are very different. It is through its association with big data technologies that the exposome approach undoubtedly distinguishes itself from molecular epidemiology. Indeed, the exposome belongs also to a context in which

140

É. GIROUX

the gathering of massive data and its bioinformatic processing has become easier. The idea is that these new technological capabilities of data gathering and processing will allow for a so-called agnostic approach, in the sense that they do not depend on prior hypotheses, as is the case in classical epidemiology (Rappaport 2011). With the exposome, then, not only is one no longer restricted to the study of some reference substances taken in isolation, but one equips oneself with the means to discover new noxious exposures. To summarize, this section has shown that the originality of the exposome concept resides in the importance accorded to the notion of the internal exposome and the replacement of the notion of environment with a very wide conception of exposure. This importance is associated with promoting the interest of “omic” biomarkers in the context of large epidemiological cohort studies. Those biomarkers are considered as being more specific and sensitive, and also more precise (Wild 2009, 119). Beyond this common core, one observes in the literature two main orientations in the interpretation of the exposome and its integrative dimension. On the one hand, it is considered that the exposome consists essentially in the study of internal exposures which appear to occupy the role of interface between the external and the internal and which improve the level of resolution of exposure measurements and the explanation of mechanisms linking exposure to disease. On the other hand, more attention is given to the multiplicity of sources, types and levels of exposures, and to the holistic, comprehensive and highly integrative nature of the exposome. The two following sections offer successive examinations of these orientations and their respective conception of the integrative dimension of the exposome.

3 3.1

The Expos-omic: The Centrality of the Internal Exposome

A Specific View of Precision: Centrality of the Omics Approach

It is important to come back to the post-genomic context of the introduction of the concept of the exposome by Wild in 2005. The aim is that of “complementing the genome”. There is continuity here with the Human Genome Project. The latter raised significant hopes in terms of better understanding the aetiology of chronic diseases. But these hopes met with disappointment from the results of genome-wide association

A CRITICAL ASSESSMENT OF EXPOSURES INTEGRATION …

141

studies (GWAS): the associations are too weak to be conclusive; the studies fail to account for the bulk of variability of these diseases (Lioy and Rappaport 2011). The “Human Exposome Project”8 aims to make up for these insufficiencies and at long last to provide the means for obtaining a complete knowledge of disease aetiology. The objective is to complete the genome-wide association studies with exposome-wide association studies (EWAS). As underlined above, this link with the genome inscribes the concept of the exposome in the wake of genetic or molecular epidemiology. Moreover, in addition to the concern for greater precision, the importance accorded to collecting large amounts of data establishes a continuity with “precision medicine”.9 The vocabulary of “precision” is indeed omnipresent in the literature on the exposome. The motive repeatedly affirmed as lying at the origin of the field is the improvement of the precision of environmental measurements in order to obtain the same level of resolution as that obtained for the genome in the context of sequencing (Lioy and Rappaport 2011). But greater precision seems immediately to imply, and without this really being examined, the biological and even “omics” characterisation of an exposure. For some, the exposome is nothing other than “an omic-scale characterisation of the nongenetic drivers of health and disease” (Niedzwiecki et al. 2019). The precision here is then almost equated with the “omics” approach and the concept of exposome is thus fully and explicitly assimilated or reduced to the expos-omic.10 But other authors like Wild affirm that the exposome is not reducible to the omics characterisation of exposures. First, one must distinguish between the measured phenomenon (exposome) and the method (exposomic) by which it is apprehended. And secondly, this method is itself not the only one: other methods, such as geo-spatial 8 https://humanexposomeproject.com. 9 See footnote 2. 10 The meaning attributed to the suffix “ome” in “exposome” is ambiguous. It can just as well designate the totality of exposures, whatever their nature might be, as the totality of certain sorts of biomarkers of exposure at the molecular level. In this latter case, “ome” has the meaning of the suffix used in so-called “omic” technologies, which concern the systematic analysis of the totality of certain sorts of molecules: DNA (genomic); RNA (transcriptomic); proteins (proteomic); cellular metabolites (metabolomic); and lipids (lipidomic). For greater clarity, it would be helpful to speak only of the “exposomic” in this latter sense (I use the notation “expos-omic”) and to privilege the notion of “exposology” for the former.

142

É. GIROUX

or personal monitoring, mobile telephones, etc., are described as useful (see Wild et al. 2013, 482). Nevertheless, the “omics” technologies have a special role here because the attractive promise they make is “the capture of a wide-range of exposures in a single measurement” (Wild et al. 2013). In this context, one readily understands that the successive definitions of the exposome, initially conceived in a large and global sense, have placed the internal exposome at the centre and accorded it a privileged status. It is this exposome that underpins the promises of innovation and knowledge. It promises to improve the knowledge of mechanisms linking external exposure to internal exposure and then to disease and holds the key to precision. Moreover, the internal exposome contain traces of the external exposome and could then play the role of synthetic indicator of the totality of the exposome. It thus seems to be assumed that studying the internal exposome could be sufficient to understanding the whole exposome. A specific illustration are the hopes raised for the analysis of the “blood exposome” (Rappaport et al. 2014). Indeed, Rappaport (2011) particularly insists on the internal chemical environment of the body and what he calls a “top-down” approach based on the non-targeted study of blood samples at multiple key moments in life, in opposition to the dominant “bottom-up” approach of traditional epidemiology, focussed on the study of air, water, diet, etc., i.e. the external exposome. It may be observed that the choice of the expression “top-down” for this way of apprehending the environment is surprising, for it is usually associated rather with a holistic or emergentist approach, and “bottom-up” approaches being associated rather with reductionism. This adds to the confusion between the holistic and the more reductionist orientations of expos-omic research. So, in this conception of the exposome, the body is itself perceived as an environment in which diverse chemical substances or endogenous processes are considered as exposures. This allows epidemiology to study the internal components of exposure, that is to say, what the external exposure initiates as an event in the individual organism, which may be identified and measured in a precise way by means of molecular biomarkers. Furthermore, the external exposome, and the social factors in particular, but also and more generally the sources of exposures, find themselves relegated to the background. This is true to the point that certain researchers have introduced the notion of “public health exposome” in order to serve as a reminder of the importance of the general external exposome and of completing an individual perspective

A CRITICAL ASSESSMENT OF EXPOSURES INTEGRATION …

143

by a populational perspective (Juarez et al. 2014; see also “From outside In – integrating external exposures into the exposome concept”, Cui and Balshaw 2019). But let’s turn now to an analysis of the first generation of studies which began at the beginning of the 2010s. 3.2

Primarily “Expos-omic” Studies

The first generation of studies concentrated on the analysis of biomarkers associated with certain physical environmental exposures, without looking at the social environment. Arising in the context of important international collaborations, these projects, like the majority of research in epidemiology, essentially adapted their methodology as a function of the availability of exposure data. For example, the HELIX (Human Early Life Exposome) cohort, initiated in 2013, which is a consortium of European cohorts on neonatal health bringing together 31,472 women from 9 regions and 6 countries, aims to describe the multiple environmental exposures of pregnancy and childhood. The goal is to identify associations between these early exposures, “molecular signatures” and childhood diseases (Vrijheid et al. 2014). Exemplifying the multifactorial (or synchronic) approach to exposome research, HELIX integrates the measurement of 17 different exposures (climate, air pollution, built environment, chemical agents, etc.). But HELIX only includes those geographical areas (regions, urban areas, etc.) for which information on air pollution and the built environment are available (Maitre et al. 2018) and is limited to the study of exposures at the individual level. Another project, EXPOsOMICS (2012–2017), financed by the European Union and bringing together 13 European and American research centres, aims to develop a new approach to the evaluation of exposure to environmental pollutants, by characterizing the external and internal components of the exposome. The study “focussed on two high priority environmental pollutants, air pollution and water contaminants (…) in studies of critical life stages” (Turner et al. 2018). The researchers do not, however, seem to have considered the exposome as a favourable opportunity to put in place a more multilevel and multidimensional apprehension of exposures: “on a basic level, exposome research can be seen as replicating the approaches of classic risk assessment with higher resolution and

144

É. GIROUX

greater accuracy” (Turner et al. 2018). The allegedly “holistic” or integrative dimension is contained in the taking account of several exposures at a time, regardless of the diversity of the nature and type of these exposures. By concentrating on certain environmental pollutants, these two studies apply a narrow vision of the exposome. The objective is to identify biomarkers and to focus on critical periods in life. This resembles an in-depth application of traditional research methods in environmental and molecular epidemiology. Above all, the epidemiological studies, particularly in the context of an exposomic approach, must adapt to variability in the precision of the measurements and of the measuring instruments. Whereas they benefit from the greatest precision in the biological measurements, researchers encounter greater difficulties in the routine collection of precise data about the socio-demographic and residential trajectories of individuals, but also about their current and past exposure to different agents. This imbalance leads one to give prominence to research into biomarkers and the internal exposome, which could be more easily and more precisely measured. It can indeed lead researchers, out of a concern for methodological homogeneity and optimisation of the scientific quality of their research, to only take an interest in external exposures that are translatable into, and measurable by, metabolic or epigenetic changes in the internal exposome.11 But what about the context of these exposures or more generally the environmental and socio-economic determinants which are known to have a major influence on health variations and inequalities of exposures at the population level and which are not always measurable at the individual level (Marmot et al. 1978; Marmot and Wilkinson 2005)? Behind the multidimensional and integrative intent of the exposome, it is clear that the effective application of its multidimensionality is not easy and that it requires special effort and scientific commitment to this end. But at the same time, as the geographer Prior and his co-authors have emphasized, “(t)he lack of the social is damaging to exposomic studies; environmental exposures and their biological correlates cannot be separated from the broader social, economic, political and cultural relations in which they are embedded” (Prior et al. 2019).

11 Part of those three paragraphs has been written by Yohan Fayet in the context of our collaboration in the project “EPIEXPO”. Many thanks to him.

A CRITICAL ASSESSMENT OF EXPOSURES INTEGRATION …

3.3

145

Criticisms of the Reductionism of Expos-omic Research

Social science researchers have denounced the reductionist character of the exposome in the sense that the social conditions that determine exposure variations at the population level are reduced to differences at the molecular level (Shostak and Moinester 2015). This criticism of the reductionism of the exposome joins up with those already formulated with respect to epigenetics and its ambition to integrate the environment in a holistic perspective (Niewöhner 2011; Landecker 2011; Shostak 2013, 202). This echoes the more general criticism of the biomedical paradigm, judged too pervasive in epidemiology, made by certain social epidemiologists: this paradigm leads to a focus on the individual and the biological to the detriment of the populational level and of social determinants of health (McMichael 1999; Krieger 1994). Indeed, a debate divided epidemiology in the 1990s. And the proposal for a perspective that integrates multilevel exposures had already been made in the context of a criticism of the way molecular epidemiology (Krieger 1994; Susser 1999) aimed to overcome the limits and difficulties of risk-factor epidemiology or “black-box” epidemiology (Vandenbroucke 1988). At one side, advocates of molecular epidemiology consider that the molecular level holds the key to causal explanation and to the control of disease. At the other side, epidemiologists concerned with the importance of social factors and the population level of analysis and action contributed to the institutionalisation of social epidemiology (Berkman and Kawachi 2000; Krieger 2001).12 The expos-omic, as it is described in this section, is an extension of the molecular orientation in epidemiology. The integration its advocates promote amounts to taking account of the interaction of genes and the environment, essentially at the molecular level. If one takes a closer look at the interdisciplinarity in question in the first generation of exposomic studies, it does not seem to include the social sciences: it is above all a question of interdisciplinarity between epidemiologists, biostatisticians, bioinformaticians, chemists and toxicologists. In the literature, it is noteworthy that everyone speaks of integration and interdisciplinarity but that, if the latter can extend to the social sciences in certain single-author

12 This debate has been compared to the one opposing advocates of the miasmic theory (and hygienism) to advocates of the theory of germs at the end of the nineteenth century (Loomis and Wing 1990; Susser 1999).

146

É. GIROUX

articles of Wild (e.g. 2012, 30), it is more generally limited to those disciplines that have just been mentioned (Wild 2005, 1849; Rappaport 2011, 8; Wild et al. 2013, 492). Moreover, whereas the notion of holism is often associated in the history of medicine with the demand to take the (external) environmental and population-level determinants of health into account, it is restricted here above all to the idea of simultaneously taking account of a large number of exposures measured at the molecular level. So, under the cover of a wide integrative and holistic approach, the exposome would actually appear to reinforce an individualisation, biologisation and even a “molecularisation” of the environment (Shostak and Moinester 2015; Senier et al. 2017), in logical continuity with genomic and/or precision medicine. The risk is thus to neglect social, political or structural determinants that are irreducible to, or simply not measurable at, an individual or molecular level, as well as to neglect preventive populational interventions, even though they are often more efficient. As a consequence, the exposome, reduced to the expos-omic could entail, through this biologisation and individualisation, a depoliticisation of environmental health questions (Saracci and Vineis 2007; Shostak 2013; Guchet 2019). Two unquestioned presuppositions seem central here: an ontological presupposition that every external exposure having an effect on individual health must alter our internal chemical environment, and an epistemological optimism that every exposure or noxious effect is potentially measurable at this level.

4

What About the Most Integrative Perspectives?

Other approaches leave more room for the ambition of a wider concept of the exposome, as it is defined in the princeps article of Wild (2012), with its three domains. In order to demarcate their vision from the approach to the exposome described above, Paul Juarez and his coauthors propose an “integrated” and “holistic” approach, defending a “public health exposome (PHE) approach” (Juarez et al. 2014, 12870) and, in continuity with this, Laura Senier and her co-authors (2017) put forward the concept of the “socio-exposome”. These approaches aim to leave more room for the general external exposome, but also to both

A CRITICAL ASSESSMENT OF EXPOSURES INTEGRATION …

147

individual and population levels of analysis and to social science expertise alongside the biomedical sciences.13 Others adopt another strategy and seem to go further in the objective of integration by seeking to model the causal relation between the biological and the social in the production of health effects. The underlying issue here is that of the interest social epidemiology takes in the exposomic and so, in other words, of dialogue between social epidemiology and molecular epidemiology. The elaboration of the concept of “sociomarker” by analogy with the “biomarker” is at the core of this approach. In the following and final section, the focus will turn to this second type of approach, as it is embedded in the Lifepath project. This strategy is closer to the concept of the exposome described above as according a central role to the “biomarker approach” in its study of aetiology, and thus allowing us to explore the path of a conciliation between more holistic and more reductionist approaches. In particular, integration here is explicitly explanatory in a similar sense to that theorized by certain philosophers with respect to the biological sciences in particular (Brigandt 2010).14 Does this integrative approach renew the methods and the manner of modelling disease aetiology in epidemiology by allowing for effective collaboration between the disciplines concerned? If that were the case, exposome research would be an exemplary case of this quest for interdisciplinarity necessary to the health sciences (Clarke et al. 2019). Or, more modestly, does this approach constitute an opportunity for social epidemiology, and more generally for the social sciences and the importance they accord to the social determinants of disease, to find greater credibility and legitimacy in the eyes of the biomedical sciences? But if this is the case, at what cost for the social type of explanation? It is to these questions that this section will turn.

13 The PHE approach aims at integrating “information about endogenous and exogenous exposure mechanisms, processes and outcomes with mediating and moderating factors at both the individual and population health levels” (Juarez et al. 2014, 12870). 14 This explanatory mode of integration remains a less clearly determined horizon in

approaches like the “public health exposome” (Juarez et al.) or the “socio-exposome” (Senier et al.). One may no doubt speak in the context of these approaches of a mode of integration closer to pluralism or to interdisciplinarity than to integration in the strong sense of the term (see the notion of interdisciplinary success without integration [GrüneYanoff 2016]).

148

4.1

É. GIROUX

Towards a More Integrative Perspective on Disease Aetiology

Modelling the integration of the biological and the social in the study of disease aetiology is far from being new: in the domain of clinical medicine, one could mention the “biopsychosocial” model defended by George Engel (1977). Engel put this forward as a model making it possible to fix a certain reductionism in biomedicine, overly focussed on the biological aspects of disease. The idea, recently taken up again in a contemporary context by Derek Bolton and Grant Gillett (2019), is to conceive health as constituted by these three dimensions in an irreducible manner, with each dimension having its own causal influence. The objective is better to understand the way in which they interact to produce health phenomena. Closer to the context of the exposome, from the 1970s onwards epidemiologists, and more precisely social epidemiologists, have proposed, and defended the importance of, integrating diverse types of factors at diverse levels of organisation, from the molecular to the social. Mervyn Susser defended the framework of his “ecological model” and the notion of Chinese boxes (1996), and more recently, Nancy Krieger defended an “eco-social theory” as the relevant framework of this integration (2001). As mentioned above, these social epidemiologists are opposed to the reductionism of biomedicine and risk-factor epidemiology, considered overly focussed on individual biological and behavioural factors (the specific external exposome), and neglecting the social and the populational (the general external exposome). For Susser as for Krieger, epidemiology must be both molecular and social, going beyond these divisions (Susser 1999). It is thus even described as the key discipline for a multidimensional and integrative approach to health. In this critical perspective towards biomedicine, and in order to conceptualize the importance of taking account of both the social and the biological in epidemiology, Krieger accords a central role to the notion of “embodiment”. She defines it as “a concept referring to how we literally incorporate, biologically, the material and social world in which we live, from conception to death; a corollary is that no aspect of our biology can be understood absent knowledge of history and individual and societal ways of living” (2001, 672). In her conception, embodiment is based on the idea that we are, as human beings, simultaneously biological organisms and social beings. It is first of all a “construct” and a “process” whose objective is to orient epidemiological research towards promoting “not only rigorous science but also social equity in health” (2005, 350).

A CRITICAL ASSESSMENT OF EXPOSURES INTEGRATION …

149

The epidemiologist, Paolo Vineis, who led the Lifepath project, considers that the tools and concepts of the exposomic are necessary if we are properly to understand embodiment and more generally the relations of the social to the biological in disease aetiology (2018, 361).15 But what is the meaning of the concept of embodiment in the context of this study and what type of integration does it promote? In order to respond to these questions, the Lifepath project will first be presented and then followed by an analysis of the “mixed mechanisms” of health and disease proposed by the philosophers, Ghiara and Russo, in this context, based on the concept of the “sociomarker”. To conclude, a critical analysis of the bio-social integration thus envisaged will be presented. 4.2

The Lifepath Project (2015–2019)

The objective of the Lifepath project is to study the biological processes underlying social inequalities in ageing in good health (“the biology of inequalities”, Vineis et al. 2017a) on the basis of 8 cohorts in France, Italy, Portugal, Ireland, the UK, Finland, Switzerland and Australia. Ageing is envisaged as a set of processes, both social and biological. Lifepath project is a contribution to the exposome research.16 Its studies of ageing adopt a “life course” approach (diachronic dimension of the exposome). And the project extends its investigation of the social to the biological through an interdisciplinary approach (synchronic dimension), thanks to the use of “omic” technologies and epigenomic and metabolic biomarkers. It thus includes the three principal dimensions of the exposome mentioned above. Moreover, three layers of study are tackled, which it may be noted cover the three domains of the exposome, as conceived by Wild: “outer: determinants” (general external exposome), “intermediate: risk factors” (specific external exposome) and “inner: biomarkers and omics” (internal exposome) (Vineis et al. 2017a, 425). These three layers also respectively correspond to three branches of epidemiology: social epidemiology, risk-factor epidemiology and molecular epidemiology. The originality of Lifepath with respect to HELIX or EXPOsOMICS, which are centred on the impact of certain environmental pollutants, is to 15 A reference is made to Nancy Krieger (2005). 16 That is explicit on the internet site of the study, where it is specified that Lifepath

project contributes to the development of the study of the exposome. See: https://www. lifepathproject.eu/content/exposome-new-frontier-environmental-research.

150

É. GIROUX

focus primarily on the biological effects of the socio-economic environment. The objective is to make possible the development of preventive clinical actions and to provide evidence for policies in public health seeking to tackle the problem of social inequalities in ageing and the social determinants of health. The notion of embodiment is employed to theorize the integrative approach of the biological and the social (Vineis et al. 2020, 34). The aim is to develop the means to measure the embodiment of the social, notably via the allostatic load17 and the epigenetic clock (age acceleration). The integration concerns variables and disciplines: social science and public health approaches are integrated “with biology (including molecular epidemiology)” (Vineis et al. 2017a, 418) that is to say, social epidemiology with molecular epidemiology and exposomic research. However, as in the case of the studies discussed above, such a project is not without methodological constraints when it comes to integrating biological and social variables. The cohorts of the Lifepath project are “only a small proportion of all cohorts available in Europe” because they had to combine “good measures of socioeconomic status, risk factors for non-communicable diseases (NCDs) and biomarkers already measured (or availability of blood samples for further testing)” (Vineis et al. 2017a, 418). Consequently, the variables used to measure social exposure and more specifically socioeconomic status, remain approximate variables or proxies, which are relatively imprecise: the level of education, the profession or the revenue. There does not seem to have been an investigation into the improvement of the social metrics themselves. In fact, the study uses pre-existing data for the cohorts. Particular attention is nevertheless accorded to the “stressful conditions that may be associated with social position measured using variables capturing adverse childhood conditions, and measures of socio-economic hardship, such as exposure to the great recession, or material deprivation” (Vineis et al. 2020, 13). The results of this study show that socioeconomic factors explain a large part of the variation in ageing that known classical risk factors (behaviour, lifestyle, etc.) fail to explain. In this sense, social variables are not included simply as “parasite variables”, such that they may be treated as potential factors of confusion (Bauer 2011, 307), but rather 17 The “allostatic load” represents the cumulative effects of a dysregulation of the biological system with persistent badly regulated allostatic responses. On this concept, see (Beckie 2012; Johnson et al. 2017).

A CRITICAL ASSESSMENT OF EXPOSURES INTEGRATION …

151

as variables studied for their own aetiological role. Above all, this study aims to show the value of using the “biomarker approach” for better explaining the relations between social and biological factors and health effects. It is here that the direct continuity with the exposome approach introduced by Wild is visible: the extension of the use of biomarkers to the whole of epidemiology, and in this instance to social epidemiology. The Lifepath studies show that the impact of social-economic conditions on health, and in particular on premature ageing, “is mediated by known behavioural and clinical factors” and “intermediate molecular pathways (…) including epigenetic clocks (age acceleration), inflammation, allostatic load and metabolic pathways”. These pathways “highlighting the biological imprint (embodiment) of social variables and strengthening causal attribution” (Vineis et al. 2020, 9). The idea is that biological markers of social embodiment are “intermediate mechanisms”: in this way, the study allows for progress towards an explanatory integration of the biological and social factors of disease. Let us now try to clarify this point, as well as the causal model that underpins this approach. 4.3

Sociomarker and Mixed Mechanisms

In the context of the observation of the Lifepath project and its objective of linking the social to the biological via the identification and measure of intermediate biological markers,18 the philosophers Ghiara and Russo (2019) have developed the concept of “sociomarker”, by analogy with the biomarker, as it is used in molecular epidemiology and, more generally, in exposomics. The objective is to theorize the integration of social and biological factors in a single causal model of pathogenesis and to accord them equal importance. This proposal is presented as an extension of earlier reflections on the necessity of integrating social factors along with biological ones in our understanding of pathogenesis and thus envisaging pathogenesis as a complex “mixed mechanism” (Kelly et al. 2014). The starting observation is that if social epidemiology has already to a large extent proved the existence of correlations between socio-economic determinants and a large number of health effects at the populational level, it has nevertheless not developed a causal model which allows for 18 Lifepath project is cited as a classic example of this quest for integration and as leading to the introduction of new concepts and methods to this end (Ghiara and Russo 2019, 8).

152

É. GIROUX

an integration of biological and social mechanisms at the populational and individual levels. To Ghiara and Russo (2019), the problem is that social epidemiology settles for measuring socio-economic position by means of more or less direct indicators (the level of education for example) or to integrate it “as a proxy variable” (Bauer 2011, 309), such as for example eligibility for free school meals. But these measures have a classificatory but not an explanatory function (Ghiara and Russo 2019, 9). The objective of the concept of “sociomarker” is to give social variables an explanatory role, just like biological variables. The proposed solution is to measure the social with more precision at the individual level by developing “sociomarkers”, on the model of molecular “biomarkers”, which allow for a more accurate approach to the causal chain. This is a return to the “meet-in-the-middle” method mentioned above with respect to the expos-omic. This configuration makes it possible to combine sociomarkers and biomarkers in a single model. The sociomarker is not a particular type of measurement, and it is not necessarily different from classical social indicators. What changes is the way it is used (Ghiara and Russo 2019, 15). By the simple fact that it is positioned at the individual level as an intermediate marker linking the social and the biological, it is a less approximate and more precise social measurement. By helping to detect the key stages linking external factors (macro) to factors at the individual level (micro), the markers suggest the existence of a causal continuum between the two and help to intercept the process of mechanism. They thus help one understand how socioeconomic factors “get under the skin” and affect the biological at the individual level and not only that they do that, thus revealing where one might intervene in the causal process as a whole. The Adverse Childhood Experience (ACE)—measured on the basis of a set of indicators of traumatic or stressful events experienced during childhood—is an important indicator in Lifepath and considered a classic example of a “sociomarker”. It links a social factor at the socio-economic population level (macro) to individual biological factors (micro), like the allostatic load, itself associated to health effects. It has on the one hand been observed that ACEs are more frequent in children of low-income families, and on the other hand studies have shown that they are associated with biomarkers of inflammation and allostatic load. Thus, “ACEs are not just indicators, but sociomarkers that can help to establish how socio-economic determinants exert an impact on health through biological mechanisms” at the individual level (Ghiara and Russo 2019, 17).

A CRITICAL ASSESSMENT OF EXPOSURES INTEGRATION …

153

The sociomarker thus links the biological and the social, combining the populational and the individual levels. The idea is indeed to gain increased precision in the measure of the social and to provide a mechanistic explanation of the relation between social exposure and health effects, and not only a correlation (for more details on “sociomarker” concept, see Russo, in this volume). This continuum, and this mechanistic explanation, strengthen the argument in favour of a causal relation between the social and populational factors on the one hand, and individual physiopathology on the other. But in so doing does this approach not lead one to privilege preventive actions at the individual level, or at the level of organic inflammation, rather than at the level of populational and structural factors? It would then lead one to privilege, say, potential preventive treatments for reducing the allostatic load or organic inflammation at the individual level, rather than acting at the level of the population and structural sources of exposure. 4.4

Critical Analysis and Some Thoughts on Bio-Social Integration

The Lifepath project and the concept of “mixed mechanisms of health” constitute an advance in the objective of integrating biological and social factors in the explanation of health phenomena. This integrative approach nevertheless raises a certain number of questions which are also the occasion to propose paths for further reflection on the nature, the forms and the relevance of an integrated bio-social approach to disease causation. 4.4.1 Strong but Narrow Integration One may firstly ask what becomes of the aetiology of disease in this mixed mechanism centred on pathogenesis, that is to say, the very process of the disease. Is not the risk, when focussing on pathogenesis, that of relegating to the background the socio-economic, contextual and structural factors? A more global and complete model of disease causation which integrates both the aetiological and the pathological processes, of the sort proposed by Dammann and Smart (2019) and Dammann (2020), would no doubt hold increasing relevance. Moreover, by focussing in this manner on the social and the biological, what becomes of the psychological ? It appears indirectly through the example of the ACE as mediator of the effect of the social on the biological. So is that its place in disease causation? Would it not be more

154

É. GIROUX

appropriate to imagine a certain autonomy of the psychological with respect to the social? Moreover, in Lifepath, physical or chemical environmental factors, like the pollutants studied in EXPOsOMICS and HELIX, are absent. As a result, one finds a form of restriction regarding the range of types of external exposures taken into account. Integration would be strong or intensive in being explanatory, but narrow in the range of types of exposures taken into account. Nevertheless, this is no doubt a necessary compromise between globality and precision, which may be compared to the compromise highlighted by Levins (1966) between generality and precision for scientific modelling in biology. One may thus consider that this study constitutes a first step towards a combined analysis of measurements of the physical, biological and social environments. A second generation of more broadly integrative projects, combining different types of data and measurements, and a wider diversity of studied exposures are now emerging, in the context of the European Human Exposome Network. Nevertheless, important statistical, methodological and conceptual developments are required if this path is to be pursued and the integration of data, methods and explanatory models made easier. It also seems difficult to think that the exposomic will avoid the difficulties associated with what it has become common to call the “curse of dimensionality”. It is not certain that it will be possible—or pertinent— to model disease aetiology by means of an integration that would be both strong (from an explanatory perspective) and broad (with respect to the range of types of variable covered). 4.4.2 A Specific Concept of Causation The Lifepath project, like the exposomic and the mixed mechanisms proposed by Ghiara and Russo (2019), is based on a commitment to a specific conception of the causal mechanism. But this conception is not entirely consensual in the philosophical debate about the notion of mechanism and is not necessarily the most suitable way to apprehend social mechanisms. Indeed, if a relatively consensual definition has emerged of the mechanism as a phenomenon consisting in entities or parts whose activities and interactions are organized in such a way as to be responsible for a specific phenomenon (Machamer et al. 2000), there exist several interpretations of the nature of these activities or interactions. In Lifepath and the “mixed mechanisms” approach, the mechanism is conceived by reference to the causal theory of “mark transmission” developed in the

A CRITICAL ASSESSMENT OF EXPOSURES INTEGRATION …

155

philosophy of science by Hans Reichenbach and then taken up by Wesley Salmon (1984) in the context of the physical sciences (Vineis et al. 2017b). In this theory, a causal mechanism is present when one can detect “mark transmission” between the cause and the effect and thus observe a continuum. Drawing on this, exposomic research would make it possible to identify the mechanisms that link exposure to disease in the sense that they make it possible to reconstitute a more or less continuous succession of “marks”, thanks to the identification of “intermediate biomarkers”. But this conception of mechanism gives rise to several questions. First, it is not obvious that this conception, first developed in the context of physical science (Salmon 1984) and based on an ontological vision of transmitted “marks” as physical quantities (energy for example), is applicable to the living. Nevertheless, it has been adapted to social and biological realities by Ghiara and Russo by considering, on the one hand, that what is transmitted is information (Illari and Russo 2016), and by conceptualizing markers as signals of this transmission at the level of epistemology, not of ontology: “scientists reconstruct the continuum between factors at different levels” (Ghiara and Russo, 2019, 13). It remains the case that this approach to the causal mechanism in terms of “mark transmission” was initially developed to understand physical phenomena, and that social sciences use other conceptions of mechanism that may be more relevant when tackling the social causes of disease. They may draw notably on the counterfactual conception, as put forward by Glennan (2011) or Woodward (2011). Stated briefly, this conception rests on the idea that a causal chain may be considered in terms of counterfactual dependence: C causes E if and only if, if C had not occurred, E would not have occurred. François Claveau (2012) contests the idea that the counterfactual-manipulationist account of causality19 would only give access to difference-making evidence, as Russo and Williamson would appear to suppose (2007): he shows that it can just as well supply mechanistic evidence (Claveau 2012, 810). In this counterfactual conception of mechanism, it isn’t necessary to be able to account for the detail of the path linking exposure to pathological effect and to presuppose a continuum.

19 The manipulationist concept of causation is based on the following guiding idea: the C variable is in a causal relation with respect to the E variable if and only if interventions on C make it possible to modify the value of E.

156

É. GIROUX

On the contrary, such a continuum or linear unidirectional relation is very much present and even presupposed by the conception of the mechanism as “mark transmission” and in the “meet-in-the-middle” approach. But is this not a costly supposition given the great complexity of relations and interactions between the different factors and levels involved in the production of health effects? To apply this linear vision of causation could restrict in advance the causal role given to the social and the psychic. The social only seems to be causal if it alters the biological via the psychic in a unidirectional social-psychic-biological continuum. Similarly, as underscored above, the psychic would only have a mediating role between the social and the biological, with the social remaining in the position of a “distal” factor to which one only attributes a causal role if it alters the biological individual in an identifiable and measurable way. In Lifepath, the conceptual model of embodiment over the life course is explicitly reformulated in relation to “the social-to-biological processes involved in healthy aging” (Vineis et al. 2020, 12). That is to take a step away from integration, as it is envisaged in the biopsychosocial model presented above. Accounting for interactions between different types of exposures and a causality specific to the social—that is to say, not mediated by the psychic or not necessarily translating into a biological alteration at the individual level—is largely absent, even if it would appear to be a possible and desired horizon for these approaches. Furthermore, the causal relation between individuals and the social structure or context can be recursive in nature: for example, the environment of the neighbourhood influences individual social capital and this in turn shapes the attributes of this same environment. Moreover, it is far from certain that identifying intermediate biomarkers and reconstituting a continuum makes it possible, and indeed suffices, to give a mechanistic type of evidence, as has been claimed. Stefano Canali (2019), in an epistemological analysis of the EXPOsOMICS study, showed that molecular data are in reality used as difference-making evidence. This research identifies exposure profiles and establishes lists of molecular elements and components present in the study sample but develops no knowledge of activity or organization between the molecular entities, no knowledge of how. These components are envisaged in a static manner. It is a relation of dependency between external toxic substance and internal molecular marker that is highlighted. This is closer to correlation than to mechanism. One could

A CRITICAL ASSESSMENT OF EXPOSURES INTEGRATION …

157

perhaps consider that this provides clues to the identification of mechanisms, but there is in any case no production of mechanistic evidence. For that, it is necessary to draw on other types of method and knowledge. 4.4.3 A Specific Conception of Causal Inference and of Evidence Further, the conception of causal inference in Lifepath and the mixed mechanisms approach is based on an evidential pluralism mentioned in Sect. 2.1, which defends the necessary presence of two types of evidence. The isolated identification of a mechanism or of a correlation (differencemaking evidence) is insufficient. Both types of evidence would be required to establish a causal claim in medicine (Russo and Williamson 2007; Vineis et al. 2017b). This conception is used to justify, in the context of exposomics, the importance accorded to the fact of complementing the evidence of correlation arising in risk-factor epidemiology by the identification of intermediate molecular biomarkers making it possible to bring to light the mechanisms occurring between the exposure and its effect. But this evidential pluralism is the subject of debate in philosophy of medicine. Some consider that it is not always necessary to have these two types of evidence in order to establish causal claims (Howick et al. 2013; Broadbent 2011). Beyond the problem of spelling out what one means by one (mechanism) and the other (correlation or difference-making evidence) (see Illari 2011), Broadbent (2011) has called into question the interest of establishing this dualism as a methodological or normative principle. He showed clearly that the mechanism certainly has an interest for explaining a correlation, but it is not necessarily a reliable guide for causal inference in epidemiology. In this discipline the objective is indeed to identify the relevant causal relation for action. In other words, it is not always necessary to identify the underlying mechanism if we are to consider that the hypothesis is causal, even if the mechanism is required for the relation to be explained. I concur with this approach and with a pragmatic and contextual conception of evidence and causal inference of the sort defended by Julian Reiss (2015, 2016) or Anya Plutynski (2018): the evidence needed to establish a causal claim is inextricably dependent on the context and objectives of the research. This is all the more important to recognize in the context of medicine, epidemiology and public health, where causal analysis is tightly linked to decisions to act and intervene: the aim is not only to understand the world, but also to change it. Consequently, more so than maintaining the necessity of obtaining at least one piece of evidence

158

É. GIROUX

for each of the two main types (mechanistic and difference-making), it seems that what matters is above all the variety and independence of the evidence, regardless of its type; the level and nature of the evidence required depends on the context and objective of the research question (see also Claveau 2012). To summarize, the above considerations relativize the weight of evidence accorded, in the field of exposomics, to the identification of the details of biological (and even molecular) mechanisms, conceived as mark transmissions at the “omic” level, through which social exposures have an effect on health. And they lead one to question the relevance and central place accorded to the mechanism in Ghiara and Russo’s model of pathogenesis. Other types of evidence and other evidence from other types of causal relations may be of greater relevance in the medical context in which intervention is the ultimate horizon. 4.4.4 Place of the Social Sciences In fine, the solution put forward for improving the integration of the social and the biological in pathogenesis resembles an alignment of the methodologies of the social science or social epidemiology with those of biomedicine and molecular epidemiology. It is based on the idea that the mechanistic model of the biomarker (a model in which the mechanism is itself envisaged as a process of mark transmission) is the most appropriate for better apprehending the relation between exposure and health effect and for integrating factors as varied as air pollution and socio-economic context. As we have seen, there would not appear to have been, in the context of Lifepath, an investigation into the improvement of social metrics as such, of the characterisation of the context of exposures, and of the social or ecological environment itself. It is essentially on the measure of the individual biological effects of social exposures that attention has focussed. The quest for precision in the measurement of social exposures is conceived as the development of the identification and measure of biomarkers of the social, and even of “omic” biomarkers of the social. Lifepath’s special report expresses this as follows: “(d)evising an exposome approach including social factors relies on the characterisation of how social experiences are translated into specific biological alterations, using available biomarkers, including high throughput omics profiles” (Vineis et al. 2020, 20). In the end, the risk is that of only considering the social

A CRITICAL ASSESSMENT OF EXPOSURES INTEGRATION …

159

as an important dimension of health if it leaves a measurable biological trace in the individual organism and if that trace may be placed on a continuum. Although this approach claims not to be reductionist, one can question the extent to which social determinants fit the mould of this “biomarker” approach without losing their own explanatory value. Expertise in the social sciences, environmental sociology or environmental justice doesn’t really have a place here. Integration is limited to a linear conception of the causal relation between social and biological factors. The risk of the expos-omic approach arises again: by focussing on toxic effects and their inventory, it overlooks and even omits the characterisation of their source, the social environment itself and the context of the exposures, offering instead an a posteriori record of the biological changes caused by these exposures. The individual factors are clearly identified but the structural factors which produce and maintain them, the social, political and economic forces which created vulnerability to exposure and are the “fundamental causes” (Link and Phelan 1995) of health inequalities, are left out of the picture. Options for collective social interventions thus risk being omitted from medical and scientific discourse. But for many social epidemiologists drawing on Geoffrey Rose’s distinction (1985) between the “cause of causes” (or “cause of incidence”) and the “cause of cases”, causation at the populational level (cause of incidence) is not reducible to what may be observed at the individual level. It is not possible, for example, to reduce social inequalities to a series of factors acting at the individual level (Giroux 2021c). 4.4.5 Which “Embodiment” Are We Talking About? As a result of the relative place accorded to social sciences in this integrative approach, the way embodiment is conceived differ somewhat from Krieger’s conception. In Lifepath, this notion actually seems to refer solely to a biological phenomenon. At issue here are markers of social embodiment (Vineis et al. 2020, 23) and social embodiment is defined « as the sustainable biological response to social adversity and related chronic stress. It involves the (possibly persistent) dysregulation of several physiological systems” (Vineis et al. 2020, 20). But for Krieger (2001, 673), the eco-social framework, in which the notion of embodiment is central, is “more than simply adding ‘biology’ to ‘social’ analyses, or ‘social factors’ to ‘biological’ analyses (…)”.

160

É. GIROUX

In fact, to speak of “social embodiment” is strange, because embodiment is both a social and a biological phenomenon. If Krieger and Smith speak of “embodied social experience” (2004, 95), it then becomes a question of a biological interpretation of the phenomena, itself socially determined (Krieger 2001, 672), and of biological phenomena, which in turn shape the social: living beings also shape their environment and do not only passively respond to it (see Krieger 2005). There are reciprocal interactions. The causal link is not unidirectional: “Embodiment is thus more than just about how social conditions ‘get under the skin’ or become embedded, as phrased in some of the contemporary psychosocial and population health literature. Embodiment instead is far more active and reciprocal, the word itself being a verb-like noun that emphasizes our bodily engagement (soma and psyche combined), individually and collectively, with biophysical world and each other” (Krieger 2011, 222). As a consequence of this, in Krieger’s conception the biological is not a pre-social universal reality, that is to say, given beforehand as fixed and identical for each individual, who responds identically to the influences of the social. The phenomenon of embodiment must always be apprehended in the historical, political and social context to which it belongs and “as a multilevel phenomenon, integrating soma, psyche, and society (…) and hence an antonym to disembodied genes, minds, and behaviours” (2005, 351). The notion of embodiment thus allows her to distinguish herself from the psychosocial theory in social epidemiology, which has paid too much attention to endogenous biological responses to stress and human interactions. She considers that in this theory what lies at the origin of these aggressions, and the way their distribution is shaped by economic and social policies, are overlooked (Krieger 2001). Moreover, even if embodiment is inscribed in individual bodies, it must be apprehended at the level of populational patterns of health, diseases and well-being. What is of interest to the “eco-social theory” she defends, is to understand why and how, in everyday life, social conditions produce a certain type of distribution of health at the level of the population (Krieger 1999, 296). But in Lifepath, even if the approach adopted aims to leave room for the populational level, a central and dominant importance is given to the individual and molecular levels. 4.4.6 Integration or Unification? In this manner, the integrative bio-social approach defended in the mixed mechanism of health and disease, although it represents an interesting

A CRITICAL ASSESSMENT OF EXPOSURES INTEGRATION …

161

and stimulating alternative to the reductionist approach to the exposomic, continues to accord primacy to the methods and concepts of the biological sciences over the social sciences. Does integration necessarily require that one of the integrated disciplines gives up its own specific methods? If this is the case, is it not rather a unification than an integration? That a strongly integrative aim comes close to a form of unification and reduction, by means of a transfer of methods and approaches from one domain (molecular epidemiology) to another (social epidemiology), is not in itself problematic and may be both pertinent and heuristically fruitful, as may an approach deriving from methodological reductionism in science. The problem is rather the gap between the holistic discourses associated with the claim to deliver a wide and complete integrative vision of the phenomena based on an interdisciplinary approach, on the one hand, and the more reductionist reality of the scientific models and practices, on the other. In particular, what is at stake is to work out which advances in our knowledge of disease aetiology will open up the possibility of choosing the most effective therapeutic and preventive interventions at the individual and population levels. In the context of Lifepath, the concern is well expressed as avoiding a form of reductionism and adopting an approach that takes an interest in causal relations at different levels of organization with a view to achieving the practical end of reducing health inequalities (Vineis et al. 2017a, 418, 427). Paths tending towards a multilevel approach are also evoked by Ghiara and Russo (2019) as a possible horizon for their mixed mechanisms approach. But if this approach aims to make room for social determinants within a combined vision of the social and the biological, the socio-bio-marker approach adopted seems to lead to a focus on those specific determinants whose precise biological impact on individuals could be demonstrated and to privilege the biological and individual level of intervention. The risk is therefore that of overlooking other more structural or contextual determinants of health and disease, like for example the processes of political regulation which authorize or perpetuate the production of risks and inequalities of exposure. Their identification depends on expertise in the historical, social or political sciences. Acting at the level of the population, these determinants are often not measurable at the individual and biological level. Above all, it is often interventions at this population level, even when these determinants, “cause of causes” (Rose 1992) or “fundamental causes” (Link & Phelan 1995), have a

162

É. GIROUX

measurable effect at the individual level, which are the most relevant and the most efficient. It remains the case that this mixed mechanisms approach, like the Lifepath project, possess originality and interest thanks to the way it promotes the integration of social factors alongside biological factors with an equal epistemic relevance in the modelling of disease aetiology. The main issue is then how to respect the diverse ways in which the social and the biological can be relevant in a context in which the social is rarely seen or accepted as being causal in itself. They also have the strength of eliciting further reflection on the best ways to integrate exposures for the understanding and prevention of disease.

5

Conclusion

Exposome research now holds an important place in North American and European research-financing policies. It promises greater precision in the measurement of environmental exposures and in knowledge of their role in disease aetiology, by integrating genetic and environmental factors. The approach adopted in this contribution has been to take seriously the scientific commitment to this research orientation, attempting to obtain a better understanding of what is meant by “exposome” in the scientific community, and, in particular, of the promises of an integrative approach to exposures. A tension is observed between two tendencies. When research concentrates on the internal exposome, the study of the external exposome becomes restricted and dependent on what may be identified at the internal level, which thus becomes the pivot and heart of a more precise consideration of several exposures at a time. Exposome research thus affiliates itself with an extension of molecular epidemiology and a strengthened integration of this discipline with toxicology and the exposure sciences. The integration and interdisciplinarity at issue here are restricted to the biological and statistical sciences. This approach appears to be far away from the often formulated ambition of developing a holistic, general and comprehensive science of exposure. When exposome research extends to social exposures and aims to establish a mixed bio-social approach of the mechanisms of disease, the tendency is to adopt the model of the biological sciences, in particular the biomarker approach, to apprehend more precisely the biological embodiment of social exposures at the level of the individual organism.

A CRITICAL ASSESSMENT OF EXPOSURES INTEGRATION …

163

We have shown that in both cases, the risk is that the demand and the quest for precise measurements, which is easier to achieve at the level of the internal exposome, prevails over the question of the pertinence and the significance of exposures in the health of individuals and populations. The risk is thus losing sight of the specificity of the social and of the population level of analysis. One touches here on the difficulties present in the relation between the human and social sciences and the natural sciences as they seek for a stronger dialogue in their approach to environmental exposures and their effects on human health (see Cavalin and Chiapperino, in this volume). One may nevertheless ask whether a pluralistic approach which shows greater respect for the specificity and expertise of each discipline, and for their own concepts and methods, at the cost of weaker integration, would not be more suitable to a comprehensive and holistic study of exposures in health, in particular when linked to a public health aim of reducing health inequalities. It is thus far from certain that the exposome will facilitate the integration of the social and molecular branches of epidemiology, or that it will realize the integrative project of eco-social epidemiology, but it nevertheless open up paths for integrating social factors in the explanation of disease in the context of epidemiological studies and methods anchored in the biological sciences (see Russo in this volume). Such research contributes substantial explanations of the mechanisms by which certain social factors act on the biological health of individuals. But more generally, it calls above all for further reflection on the type and intensity of integration that is most relevant for apprehending the causation of complex chronic diseases and facilitating their prevention at both the individual and population levels. Acknowledgements I would first like to thank Henry Dicks for his precious linguistic help. I also warmly thank the two blind reviewers for their very judicious remarks and advices. This contribution is the fruit of work carried out in the pluridisciplinary research programme “EPIEXPO” (“For a critical epistemology of the exposome”) at the Rabelais Institute in Lyon (ANR-17-CONV-0002 PLASCAN grant). Yohan Fayet and Thibaut Serviant-Fine have made important contributions to the development of the present reflections. I warmly thank them for our close and constructive collaboration in this context. I would also like to thank Séverine Louvel, Marc Billaud, Pierre-Olivier Méthot, Federica Russo and Francesca Merlin for their insightful comments on an earlier version of this manuscript.

164

É. GIROUX

References Barker, David J., C. Osmond, J. Golding, D. Kuh, and M.E. Wadsworth. 1989. Growth in utero, blood pressure in childhood and adult life, and mortality from cardiovascular disease. British Medical Journal 298: 564–567. https:// doi.org/10.1136/bmj.298.6673.564. Bauer, Susanne. 2011. Appréhender le social dans la recherche épidémiologique. Vers une politique des associations statistiques? In De la vie biologique à la vie sociale. Approches sociologiques et anthropologiques, eds. Janina Kehr, Jörg Niewhöner et Joëlle Vally, Paris: La Découverte, 298–327. Beckie, Theresa M. 2012. A systematic review of allostatic load, health, and health disparities. Biological Research for Nursing 14: 311–346. https://doi. org/10.1177/1099800412455688. Berkman, Lisa F., and Ichiro Kawachi. 2000. Social Epidemiology. Oxford: Oxford University Press. Bolton, Derek, and Grant Gillett. 2019. The Biopsychosocial Model of Health and Disease: New Philosophical and Scientific Developments. Palgrave Pivot. Brigandt, Ingo. 2010. Beyond reduction and pluralism: Toward an epistemology of explanatory integration in biology. Erkenntnis 73: 295–311. Broadbent, Alex. 2011. Inferring causation in epidemiology: Mechanisms, black boxes, and contrasts. In Causality in the Sciences, edited by Phyllis McKay Illari, Federica Russo and Jon Williamson. Oxford: Oxford University Press, 45–69. https://doi.org/10.1093/acprof:oso/9780199574131.003.0003. Buck Louis, Germaine M., Melissa M. Smarr, and Chirag J. Patel. 2017. The exposome research paradigm: An opportunity to understand the environmental basis for human health and disease. Current environmental health reports 4 (1): 89–98. https://doi.org/10.1007/s40572-017-0126-3. Canali, Stefano. 2019. Evaluating evidential pluralism in epidemiology: Mechanistic evidence in exposome research. History and Philosophy of the Life Sciences 41 (1): 4. https://doi.org/10.1007/s40656-019-0241-6. Clarke, Brendan, Virginia Ghiara, and Federica Russo. 2019. Time to care: Why the humanities and the social sciences belong in the science of health. BMJ Open 9: e030286. Claveau, François. 2012. The Russo–Williamson Theses in the social sciences: Causal inference drawing on two types of evidence. Studies in History and Philosophy of Science Part C: Studies in History and Philosophy of Biological and Biomedical Sciences (43) 806–813. Collins, Francis S., and Harold Varmus. 2015. A new initiative on precision medicine. The New England Journal of Medicine 372: 793–795. https://doi. org/10.1056/NEJMp1500523.

A CRITICAL ASSESSMENT OF EXPOSURES INTEGRATION …

165

Cui, Yuxia, and David Balshaw. 2019. From the Outside In: Integrating External Exposures into the Exposome Concept. In Unraveling the Exposome, A Practical View, edited by Sonia Dagnino and Anthony Macherone, Springer, 255–276. https://doi.org/10.1007/978-3-319-89321-1_10 Dammann, Olaf. 2020. Etiological Explanations: Illness Causation Theory. Boca Raton: CRC Press. https://doi.org/10.1201/9780429184147. Dammann, Olaf, and Benjamin Smart. 2019. Causation in Population Health Informatics and Data Science. Springer Nature. Engel, George L. 1977. The need for a new medical model: A challenge for biomedicine. Science 196: 129–136. Ghiara, Virginia, and Federica Russo. 2019. Reconstructing the mixed mechanisms of health: The role of bio-and sociomarkers. Longitudinal and Life Course Studies 10: 7–25. Giroux, Élodie. 2013. The Framingham study and the constitution of a restrictive concept of risk factor. Social History of Medicine 26: 94–112. Giroux, Élodie. 2021a. L’exposome: vers une science intégrative des expositions? Lato Sensu, Revue De La Société de Philosophie Des Sciences 8 (3): 9–28. https://doi.org/10.20416/LSRSPS.V8I3.2 Giroux, Élodie. 2021b. L’exposome: entre globalité et précision. Bulletin d’Histoire et d’Épistemologie des Sciences de la Vie 28 (2): 119–148. Giroux, Élodie. 2021c. Can populations be healthy? Perspectives from Georges Canguilhem and Geoffrey Rose. History and Philosophy of the Life Sciences 43 (4): 111. https://doi.org/10.1007/s40656-021-00463-x. Giroux, Élodie, Yohan Fayet, and Thibaut Serviant-Fine. 2021. L’Exposome: tensions entre holisme et réductionnisme. médecine/sciences 37 (8–9): 774– 778. https://doi.org/10.1051/medsci/2021092. Glennan, Stuart. 2011. Singular and general causal relations: A mechanist perspective. In Causality in the Sciences, edited by Phyllis McKay Illari, Federica Russo and Jon Williamson, Oxford: Oxford University Press, 315– 325. https://doi.org/10.1093/oxfordhb/9780199279739.003.0016. Grüne-Yanoff, Till. 2016. Interdisciplinary success without integration. European Journal for Philosophy of Science 6: 343–360. https://doi.org/10.1007/s13 194-016-0139-z. Guchet, Xavier. 2019. De la médecine personnalisée à l’exposomique. Environnement et santé à l’ère des big data. Multitudes 75: 72–80. https://doi.org/ 10.3917/mult.075.0072. Howick, Jeremy, Paul Glasziou, and Jeffrey K. Aronson. 2013. Problems with using mechanisms to solve the problem of extrapolation. Theoretical Medicine and Bioethics 34: 275–291. Illari, Phyllis McKay. 2011. Mechanistic evidence: Disambiguating the Russo– Williamson Thesis. International Studies in the Philosophy of Science 25: 139– 157. https://doi.org/10.1080/02698595.2011.574856.

166

É. GIROUX

Illari, Phyllis McKay, and Federica Russo. 2016. Information channels and biomarkers of disease. Topoi 35: 175–190. Johnson, Sarah C., Francesca L. Cavallaro, and David A. Leon. 2017. A systematic review of allostatic load in relation to socioeconomic position: Poor fidelity and major inconsistencies in biomarkers employed. Social Science and Medicine 192: 66–73. https://doi.org/10.1016/j.socscimed.2017.09.025. Joly, Pierre Benoit. 2010. On the economics of techno-scientific promises. In Débordements. Mélanges offerts à Michel Callon, eds. Madeleine Akrich, Yannick Barthe, Fabian Muniesa, et Philippe Mustar. Paris: Presses des Mines, 203–222. Juarez, Paul D., Patricia Matthews-Juarez, Darryl B. Hood, Wansoo Im, Robert S. Levine, Barbara J. Kilbourne, Michael A. Langston, et al. 2014. The public health exposome: A population-based, exposure science approach to health disparities research. International Journal of Environmental Research and Public Health 11: 12866–12895. https://doi.org/10.3390/ijerph111 212866. Kelly, Michael P., Rachel S. Kelly, and Federica Russo. 2014. The integration of social, behavioral, and biological mechanisms in models of pathogenesis. Perspectives in Biology and Medicine 57 (3): 308–328. https://doi.org/10. 1353/pbm.2014.0026. Kortenkamp, Andreas, Olwenn Martin, Michael Faust, Richard Evans, Rebecca McKinlay, Frances Orton, and E Rosivatz. 2011. State of the art assessment of endocrine disrupters: Final report. Brussels. European Commission. http://ec.europa.eu/environment/chemicals/endocrine/pdf/ sota_edc_final_report.pdf. Krieger, Nancy. 1994. Epidemiology and the web of causation: Has anyone seen the spider? Social Science & Medicine 39: 887–903. Krieger, Nancy. 1999. Embodying inequality: A review of concepts, measures, and methods for studying health consequences of discrimination. International Journal of Health Services 29: 295–352. Krieger, Nancy. 2001. Theories for social epidemiology in the 21st century: An ecosocial perspective. International Journal of Epidemiology 30: 668–677. Krieger, Nancy. 2005. Embodiment: A conceptual glossary for epidemiology. Journal of Epidemiology and Community Health 59 (5): 350–355. https:// doi.org/10.1136/jech.2004.024562. Krieger, Nancy. 2011. Epidemiology and the people’s health: Theory and context. Oxford: Oxford University Press. Krieger, Nancy. 2017. Health equity and the fallacy of treating causes of population health as if they sum to 100%. American Journal of Public Health 107: 541–549.

A CRITICAL ASSESSMENT OF EXPOSURES INTEGRATION …

167

Krieger, Nancy, and George Davey Smith. 2004. ’Bodies count’, and body counts: Social epidemiology and embodying inequality. Epidemiologic Reviews 26: 92–103. https://doi.org/10.1093/epirev/mxh009. Kuh, Diana and Ben-Schlomo Yoav. 1997. A life course approach to chronic disease epidemiology: Tracing the origins of ill health from early to adult life. Oxford: Oxford University Press. Landecker, Hannah. 2011. Food as exposure: Nutritional epigenetics and the new metabolism. BioSocieties 6: 167–194. https://doi.org/10.1057/biosoc. 2011.1. Levins, Richard. 1966. The strategy of model building in population biology. American Scientist 54: 421–431. Link, Bruce G., and Jo Phelan. 1995. Social conditions as fundamental causes of disease. Journal of Health and Social Behavior 80–94. https://doi.org/10. 2307/2626958. Lioy, Paul J., and Stephen M. Rappaport. 2011. Exposure science and the exposome: An opportunity for coherence in the environmental health sciences. Environmental Health Perspectives 119: A466-467. https://doi.org/ 10.1289/ehp.1104387. Longino, Helen E. 2013. Studying human behavior: How scientists investigate aggression and sexuality. Chicago: University of Chicago Press. Loomis, Dana, and Steve Wing. 1990. Is molecular epidemiology a germ theory for the end of the twentieth century? International Journal of Epidemiology 19: 1–3. Machamer, Peter K., Lindley Darden, and Carl F. Craver. 2000. Thinking about mechanisms. Philosophy of Science 67: 1–25. Maitre, Lea, Jeroen de Bont, Maribel Casas, Oliver Robinson, Gunn Marit Aasvang, Lydiane Agier, Sandra Andrusaityte, et al. 2018. Human Early Life Exposome (HELIX) study: A European population-based exposome cohort. BMJ Open 8: e021311. https://doi.org/10.1136/bmjopen-2017-021311. Marmot, Michael G., Geoffrey Rose, Martin Shipley, and Peter J. Hamilton. 1978. Employment grade and coronary heart disease in British civil servants. Journal of Epidemiology & Community Health 32: 244–249. Marmot, Michael, and Richard Wilkinson. 2005. Social determinants of health. Oxford: Oxford University Press. McMichael, Anthony J. 1999. Prisoners of the proximate: Loosening the constraints on epidemiology in an age of change. American Journal of Epidemiology 149: 887–897. Miller, Gary W. 2014. The exposome: A primer. Amsterdam and Boston: Academic Press. Miller, Gary W., and Dean P. Jones. 2014. The nature of nurture: refining the definition of the exposome. Toxicological Sciences : An Official Journal of the Society of Toxicology 137: 1–2. https://doi.org/10.1093/toxsci/kft251.

168

É. GIROUX

Mitchell, Sandra D. 2002. Integrative pluralism. Biology and Philosophy 17: 55– 70. Morange, Michel. 1998. La part des gènes. Paris: Odile Jacob. National Research Council. 2011. Toward precision medicine: Building a knowledge network for biomedical research and a new taxonomy of disease. National Academies Press. Niedzwiecki, Megan M., Douglas I. Walker, Roel Vermeulen, Marc ChadeauHyam, Dean P. Jones, and Gary W. Miller. 2019. The exposome: Molecules to populations. Annual Review of Pharmacology and Toxicology 59: 107–127. https://doi.org/10.1146/annurev-pharmtox-010818-021315. Niewöhner, Jörg. 2011. Epigenetics: Embedded bodies and the molecularisation of biography and milieu. BioSocieties 6: 279–298. Plutynski, Anya. 2013. Cancer and the goals of integration. Studies in History and Philosophy of Science Part C: Studies in History and Philosophy of Biological and Biomedical Sciences, 44 (4): 466–476. https://doi.org/10.1016/j.shpsc. 2013.03.019. Plutynski, Anya. 2018. Explaining cancer: Finding order in disorder. Oxford: Oxford University Press. Prior, Lucy, David Manley, and Clive E. Sabel. 2019. Biosocial health geography: New ‘exposomic’ geographies of health and place. Progress in Human Geography 43: 531–552. Rappaport, Stephen M. 2011. Implications of the exposome for exposure science. Journal of Exposure Science & Environmental Epidemiology 21: 5–9. https:// doi.org/10.1038/jes.2010.50. Rappaport, Stephen M. 2018. Redefining environmental exposure for disease etiology. NPJ Systems Biology and Applications 4: 30. https://doi.org/10. 1038/s41540-018-0065-0. Rappaport, Stephen M., Dinesh K. Barupal, David Wishart, Paolo Vineis, and Augustin Scalbert. 2014. The blood exposome and its role in discovering causes of disease. Environmental Health Perspectives 122: 769–774. https:// doi.org/10.1289/ehp.1308015. Rappaport, Stephen M., and Martyn T. Smith. 2010. Environment and disease risks. Science 330: 460–461. https://doi.org/10.1126/science.1192603. Reiss, Julian. 2015. Causation, Evidence, and Inference. New York: Routledge. https://doi.org/10.4324/9781315771601. Reiss, Julian. 2016. Causality and causal inference in medicine. In The Routledge Companion to Philosophy of Medicine, edited by Miriam Solomon, Jeremy R. Simon and Harold Kincaid. Routledge, 72–84. https://doi.org/10.4324/ 9781315720739-12. Rose, Geoffrey. 1985. Sick individuals and sick populations. International Journal of Epidemiology 14: 32–38. https://doi.org/10.1093/ije/14.1.32.

A CRITICAL ASSESSMENT OF EXPOSURES INTEGRATION …

169

Rose, Geoffrey. 1992. The Strategy of Preventive Medicine. Oxford: Oxford University Press. Russo, Federica, and Jon Williamson. 2007. Interpreting causality in the health sciences. International Studies in the Philosophy of Science 21: 157–170. Salmon, Wesley C. 1984. Scientific Explanation and the Causal Structure of the World. Princeton: Princeton University Press. Saracci, Rodolfo, and Paolo Vineis. 2007. Disease proportions attributable to environment. Environmental Health 6: 38. https://doi.org/10.1186/1476069X-6-38 Schulte, Paul A., and Frederica P. Perera. 1998. Molecular Epidemiology: Principles and Practices. Academic Press. Senier, Laura, Phil Brown, Sara Shostak, and Bridget Hanna. 2017. The socio-exposome: Advancing exposure science and environmental justice in a post-genomic era. Environmental Sociology 3: 107–121. https://doi.org/10. 1080/23251042.2016.1220848. Shostak, Sara. 2010. Marking Populations and Persons at Risk: Molecular Epidemiology and Environmental Health in Biomedicalization. In Biomedicalization: Technoscience, Health, and Illness in the US, edited by Adele E. Clarke, Laura Mamo, Jennifer R. Fosket, Jennifer R. Fishman and Janet K. Shim. Durham, USA: Duke University Press, 242–262. https://doi.org/10. 1515/9780822391258-012. Shostak, Sara. 2013. Exposed Science: Genes, the Environment, and the Politics of Population Health. University of California Press. Shostak, Sara, and Margot Moinester. 2015. The missing piece of the puzzle? Measuring the environment in the postgenomic moment. In Postgenomics: Perspectives on Biology After the Genome, edited by Sarah Richardson and Hallam Stevens, Durham: Duke University Press, 192–209. Siroux, Valerie, Lydiane Agier, and Remy Slama. 2016. The exposome concept: A challenge and a potential driver for environmental health research. European Respiratory Review : An Official Journal of the European Respirato Society 25: 124–129. https://doi.org/10.1183/16000617.0034-2016.zz Slama, Rémy. 2017. Le mal du dehors: l’influence de l’environnement sur la santé. Editions Quae. Susser, Mervyn. 1999. Should the epidemiologist be a social scientist or a molecular biologist? International Journal of Epidemiology 28: S1019–S1019. https://doi.org/10.1093/oxfordjournals.ije.a019905. Susser, Mervyn, and Ezra Susser. 1996. Choosing a future for epidemiology: II. From black box to Chinese boxes and eco-epidemiology. American Journal of Public Health 86: 674–677.

170

É. GIROUX

Turner, Michelle C., Paolo Vineis, Eduardo Seleiro, Michaela Dijmarescu, David Balshaw, Roberto Bertollini, Marc Chadeau-Hyam, et al. 2018. EXPOsOMICS: Final policy workshop and stakeholder consultation. BMC Public Health 18: 260. https://doi.org/10.1186/s12889-018-5160-z. Vandenbroucke, Jan Paul. 1988. Is ’the causes of cancer’ a miasma theory for the end of the twentieth century? International Journal of Epidemiology 17: 708–709. Vineis, Paolo. 2018. From John Snow to omics: The long journey of environmental epidemiology. European Journal of Epidemiology 33: 355–363. https://doi.org/10.1007/s10654-018-0398-4 Vineis, Paolo, Mauricio Avendano-Pabon, Henrique Barros, Mel Bartley, Cristian Carmeli, Luca Carra, Marc Chadeau-Hyam, Giuseppe Costa, Cyrille Delpierre, and Angelo D’Errico. 2020. Special Report: The Biology of Inequalities in Health: The Lifepath Consortium. Frontiers in Public Health 8: 118. https://doi.org/10.3389/fpubh.2020.00118. Vineis, Paolo, Mauricio Avendano-Pabon, Henrique Barros, Marc ChadeauHyam, Giuseppe Costa, Michaela Dijmarescu, Cyrille Delpierre, et al. 2017a. The biology of inequalities in health: the LIFEPATH project. Longitudinal and Life Course Studies 8: 417–449. https://doi.org/10.14301/llcs. v8i4.448. Vineis, Paolo, Phyllis McKay Illari, and Federica Russo. 2017b. Causality in cancer research: A journey through models in molecular epidemiology and their philosophical interpretation. Emerging Themes in Epidemiology 14: 7. https://doi.org/10.1186/s12982-017-0061-7. Vrijheid, Martine, Remy Slama, Oliver Robinson, Leda Chatzi, Muireann Coen, Peter van den Hazel, Cathrine Thomsen, et al. 2014. The human earlylife exposome (HELIX): Project rationale and design. Environmental Health Perspectives 122: 535–544. https://doi.org/10.1289/ehp.1307204. Wild, Christopher P. 2009. Environmental exposure measurement in cancer epidemiology. Mutagenesis 24: 117–125. https://doi.org/10.1093/mutage/ gen061. Wild, Christopher P., Augustin Scalbert, and Zdenko Herceg. 2013. Measuring the exposome: A powerful basis for evaluating environmental exposures and cancer risk. Environmental and Molecular Mutagenesis 54: 480–499. https:// doi.org/10.1002/em.21777. Wild, Christopher P. 2005. Complementing the genome with an ’exposome’: The outstanding challenge of environmental exposure measurement in molecular epidemiology. Cancer Epidemiology, Biomarkers & Prevention. 14: 1847–1850. https://doi.org/10.1158/1055-9965.EPI-05-0456.

A CRITICAL ASSESSMENT OF EXPOSURES INTEGRATION …

171

Wild, Christopher Paul. 2012. The exposome: From concept to utility. International Journal of Epidemiology 41: 24–32. https://doi.org/10.1093/ije/ dyr236. Woodward, James. 2011. Mechanisms revisited. Synthese 183 (3): 409–427.

From Exposome to Pathogenic Niche. Looking for an Operational Account of the Environment in Health Studies Gaëlle Pontarotti

1

and Francesca Merlin

Introduction

The role of the environment in the etiology of diseases, in particular non-communicable ones such as cancers, asthma and obesity, has, since the 2000s, increasingly been at the core of research studies in environmental health sciences. A new research domain called “exposomics” has emerged in this context. Exposomics has a heterogeneous background because it involves the concerns of two different disciplines: toxicology (i.e., the study of the adverse effects of chemical substances on living organisms) and (molecular) epidemiology (i.e., the study of the patterns and determinants of health and disease in human populations (cf. Giroux 2021a)). Moreover, exposomics participates in the emerging practice of

G. Pontarotti (B) · F. Merlin Institut d’histoire et de philosophie des sciences et des techniques (IHPST), Université Paris 1 Panthéon-Sorbonne & CNRS, Paris, France e-mail: [email protected] F. Merlin e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 É. Giroux et al. (eds.), Integrative Approaches in Environmental Health and Exposome Research, https://doi.org/10.1007/978-3-031-28432-8_7

173

174

G. PONTAROTTI AND F. MERLIN

personalized medicine, and explicitly aims at contributing to public health studies. In this research context, the term “environment” is rarely used to refer to environmental exposures. Rather, they fall under the term “exposome”, which refers to the totality of human environmental exposures from conception onwards (Wild 2005). More precisely, the exposome includes both the external exposures the body is subject to, as well as their internal manifestations (i.e., in the body). In short, the exposome is “anything that is not genetic” (Wild 2012) and “the environmental equivalent of the genome” (cf. Bacarelli 2019). Since the exposome has been conceived as complementary to the genome, research in exposomics is framed within the classical dichotomies of nature/nurture, genes/environment (Giroux 2021a). One of the most striking consequences of this view is the radical expansion of the spatiotemporal limits of the exposome, which equates to anything that is non-genetic. A salient question then is: are these spatio-temporal limits excessively expanded for the exposome to represent a relevant and operational concept of environment in human health studies? Our working hypothesis is that, contrary to what can mostly be found in exposomics, the environment relevant in health studies should be conceived as separate from the genome/environment dichotomy which has characterized biological and biomedical sciences throughout the twentieth century. Why so? We maintain this claim for the following reasons. First, identifying the environment with everything that is not genetic makes it a catch-all concept, one that is internally heterogeneous, hugely expansive and unlikely to be operational, in particular when it comes to integrating measurements of disparate exposures (from physico-chemical to sociocultural). Moreover, this definition encompasses the environment of the genes or the genomes, but not of the organisms or populations of organisms, which actually experience different states of health and disease. In other words, identifying the environment with anything that is nongenetic induces the subtle disappearance of the organism as a relevant theoretical object in the health sciences. It’s as if the organism could be reduced to the expression of its genes. Related to this second point, it’s relevant that the primacy of the gene with respect to other internal factors has been challenged over the last two decades, especially with regard to advances in epigenetics concerning the way gene expression is internally regulated, and with respect to environmental conditions. More generally, dichotomous accounts of developmental factors, as genetic and

FROM EXPOSOME TO PATHOGENIC NICHE. LOOKING …

175

non-genetic, have increasingly been questioned (cf. Griffiths and Gray 1994, Fox Keller 2010). Our objective in this paper is to answer the question of what constitutes the relevant spatio-temporal limits of what we call the “pathogenic environment” (i.e., the subset of environmental exposures that should be taken into account in order to understand the etiology of a given pathology and prevent it). We elaborate a concept of the pathogenic environment which falls outside of the dichotomous thinking that contrasts the environment (or exposome) with the genome. We build on this concept, taking into account thoughtful reflections in philosophy of biology on the concept of environment (e.g., Lewontin 1983). Finally, we propose that the concept of pathogenic environment is more operational than the concept of exposome, as it allows for ready measurement, precise prediction and environmental manipulation to prevent and/or to treat diseases. The chapter is structured as follows. First, we introduce the concept of exposome (2.1), and ask whether it corresponds to a new concept of environment (2.2, 2.3). In Sect. 3, we analyze a number of conceptual issues raised by the concept of exposome and highlight some of its shortcomings, namely the fact that it is an all-encompassing concept (3.1) which defines the environment of genes at the individual level, rather than organisms or populations (3.2). In Sect. 4, we draw on the way the concept of environment has been discussed in philosophy of biology (4.1) in order to build a concept of pathogenic environment which is intended to be more operational than that of exposome. Having provided a simple example to illustrate the relevance of this concept, which we call “pathogenic niche” (4.2), we confront our proposal with existing ecological approaches in epidemiology and public health studies (4.3), and finally analyze the literature on the obesogenic environment as a clear example of what should be conceived as a pathogenic niche (or, more specifically, obesogenic niche) (4.4).

2

Thinking About the Environment Through the Concept of Exposome

The concept of environment broadly designates what surrounds a specific entity (e.g., gene, genome, organism, group, etc.) (Pontarotti et al. 2022). Pervasive in biological, biomedical and social sciences, it has a structural importance not only in environmental health sciences, but

176

G. PONTAROTTI AND F. MERLIN

also in public health studies and epidemiology. Its use in epidemiology aims at assessing the role of environmental exposure in the etiology of chronic and infectious diseases within human populations and, on this basis, controlling health problems. Epidemiology traditionally defines the environment as “all that which is external to the individual human host” and divides it into “physical, biological, social, cultural, etc., any or all of which can influence health status of populations” (Last 1995). The question then is whether, and to what extent, the concept of exposome corresponds to a new and useful way of conceiving the environment in the study of disease etiology. 2.1

The Concept of “Exposome”

The concept of exposome has recently been proposed to designate the totality of environmental exposures that human individuals are subject to from conception to death (Wild 2005, 2012; Rappaport and Smith 2010; Lioy and Rappaport 2011; van Tongeren and Cherrie 2012). The aim of this conceptual proposition was to shed light on the variety of environmental factors involved across lifetimes in the origin of noncommunicable diseases, and thereby permit a more complete (global) and quantitative (precise) assessment of pathogenic environmental exposure. The objective here is in fine to shed new light on the etiology of these diseases.1 Thus, by adopting a global approach (both in space and time) and by looking for more precise measurements of exposures, the exposome is intended to overcome the limitations of those studies focused on a few environmental causes (e.g., air pollution) and specific time-points (Siroux et al., 2016, p. 126), as well as the shortcomings of a concept of environment dismissed as too vague (Rappaport, 2011). The exposome has been defined in different ways and includes exposures of various natures (for a review, see Giroux, this volume). To Wild, “the exposome encompasses life-course environmental exposures (including lifestyle factors), from the prenatal period onwards” (2005, p. 1848). It complements the genome. Moreover, it is possible to distinguish three “broad categories of non-genetic exposures” (Wild 2012): (1) the internal exposome (e.g., metabolism, hormones, gut microflora, etc.); (2) the specific external exposome (e.g., radiation, infectious agents, 1 For critical analyses of scientific changes brought by the exposome, its virtues and shortcomings, see Giroux (2021a, 2021b), this volume; Canali (2020), this volume.

FROM EXPOSOME TO PATHOGENIC NICHE. LOOKING …

177

environmental pollutants, lifestyle factors, etc.); (3) the general external exposome (e.g., social capital, education, psychological and mental stress, climate, etc.). In this respect, the exposome can be divided into three overlapping categories: “the internal environment of the body, the specific external agents to which one is exposed, and the social, cultural and ecological contexts in which the person lives their life” (Wild 2012, p. 25). Another seminal conception of the exposome is that of the toxicologists Rappaport and Smith (2010), who focus their research on the internal environment. According to their view, “the relevant environment” to be taken into account is “the body’s internal chemical environment” (Rappaport 2011, p. 6). The predominant importance of this internal environment is related to the fact that “the vast majority of toxic chemicals affect critical targets inside the body” (p. 6). It is also based on the hypothesis that the internal environment allows taking into account more elements than just chemicals coming from air and water (and notably from diet, smoking, drugs, radiation and endogenous processes) (p. 6). Here, the exposome can be precisely measured by biomarkers found in the blood (e.g., endocrine disruptors, immune modulators) (Rappaport et al., 2014), by adopting a “top-down” strategy2 (Rappaport and Smith 2010, p. 461). Other more integrated frameworks have been proposed to make sense of the concept of exposome (e.g., van Tongeren and Cherrie 2012, Miller and Jones 2014). However, Wild’s and Rappapport’s seminal conceptions—and the related bipartition between “internal” and “external” exposomes—are clearly the main influences on papers addressing the biggest projects in exposure science, notably HELIX (Human EarlyLife Exposome) (Vrijheid et al. 2014) and EXPOsOMICS (Turner et al. 2018). 2.2

Exposome = environment?

One can consider the exposomic literature subtly outlining, through the concept of exposome, a new concept of environment for the study of human health. In fact, studies dealing with the exposome are not very clear on the relation between the two concepts. Some papers suggest an 2 On the difference between top-down and bottom-up strategies in exposomics, see Giroux (2021b).

178

G. PONTAROTTI AND F. MERLIN

equivalence between the exposome and the environment. For example, Wild (2012, p. 24) seems to consider “environmental exposure” and “environment” as synonymous, and defines both as non-genetic. Accordingly, Vrijheid (2014, p. 876), just as Siroux and colleagues (2016, p. 125), translate Wild’s tripartition of internal, specific external and general external “exposomes” into a tripartition of internal, specific external and general external “environments”. Other studies suggest a partial overlap between exposome and environment. For instance, in a paper dedicated to the public health exposome, Juarez and colleagues (2014) identify exogenous exposures with four kinds of environment (natural, built, social, policy). However, the exposome at play in this paper is an “eco-exposome”, which is contrasted with the “endo-exposome” defined as “the inward effects arising from exposure on those receptors” (Juarez et al. 2014, p. 12868).3 From this perspective, the environment refers to the external part of the exposome (Siroux et al. 2016, p. 128). Some titles, such as “The Role of the Environment and Exposome in Atopic Dermatitis” (Stefanovic et al. 2021), seem to imply that environment and exposome are different. However, the difference between the two is not explicitly addressed in the article. Finally, Giroux (2021b) notes that in the context of studies regarding the exposome, the concept of environment generally tends to be replaced with that of environmental exposure. She notably refers to a study by Rappaport and colleagues (2014, p. 769) which replaced the concept of environment with that of exposome in order to include more than pollution in environmental exposures, and “encompass all endogenous and exogenous exposures”. According to this view, the concept of environment can be rejected for being vague and not encompassing enough (e.g., see Rappaport 2011). In this puzzling situation, one point is nonetheless clear: the concept of exposome stands for an empowered concept of environment. By “empowered” we mean more precise, more inclusive and more integrative, and, as a consequence, more scientific, in particular because of its quantitative and measurable aspect. This corresponds to the explicit project of exposomics (i.e., to precisely quantify the environment) and is clearly assumed in a

3 However, the glossary included in this paper states that “Public Health Exposome Framework considers the relationships of exogenous and endogenous exposures across multiple levels in four environmental domains: natural, built, social, and policy” (p. 12892, our emphasis). This suggests that the exposome (endogenous and exogenous exposure) can be equated with the environment (four environmental domains).

FROM EXPOSOME TO PATHOGENIC NICHE. LOOKING …

179

recent contribution by Gary Miller, the editor-in-chief of the new scientific journal Exposomics . In the editorial of its first issue (2021), Miller argues that biologists are left with “a rather fuzzy description” of the environment, while descriptions of genotypes and phenotypes are more precise. He adds, “Generally, E is defined as the environment, but the environment is enigmatic from a scientific standpoint and does not lend itself to a systematic evaluation of its constituent components”. Actually, the “(e)xposome is the superior definition of E and it provides the proper foil to G” (i.e., genotype) (2021, p. 1). 2.3

What is the Environment in Exposomics?

So, what are the new features of the environment in the context of exposomics? As shown above, it is assimilated into everything that is nongenetic (i.e., “non-genetic causal factors” of diseases) and is related to the project of building a complete etiology for non-communicable diseases. This characterization of the environment seems to be based on biologists’ theoretical background according to which the dichotomy between genes and environment is explanatorily relevant and sufficient in order to account for any trait (for instance, a disease). In this view, environment refers by principle to anything that is not genetic. In line with other authors (e.g., Giroux 2021a, 2021b), our hypothesis is that this theoretical background is the reason why the exposome is thought to “complement” the genome (Wild 2005 (Wild 2005; Vrijheid 2014). The exposome comes to complete the causal analysis of non-communicable diseases, namely to build an exhaustive account of the etiological landscape of these diseases. The implementation of the Human Exposome Project (https://humanexposomepr oject.com/) as a complement to the Human Genome Project, is indicative of this idea that is at the core of exposomics: the elucidation of the exposome will “help reveal the importance of the environment in our lives”, and “will help fulfill the promises of the Human Genome Project” (Miller and Jones 2014, p. 2). In this view, the conceptual resonance between “exposome” and “genome” can be analyzed as a condition for their epistemological equivalence, or epistemological parity. More explicitly, it appears as a device for conceiving of genes and environment as comparable and equally important causal factors in states of human health and disease, which should be studied with a similar degree of precision (Wild 2005; Siroux et al.

180

G. PONTAROTTI AND F. MERLIN

2016), and whose interactions should be better understood (Rappaport and Smith 2010). It is worth noting, for instance, that Environmental Wide Association Studies (EWAS) (Patel et al. 2010; Johnson et al. 2017) are conceived on the model of Genome Wide Association Studies (GWAS) and intended to complement them (Miller and Jones 2014). They are referred to as “the environmental equivalent of a GWAS” (Rappaport and Smith 2010, p. 461) and rest on the same kind of methodology. “The EWAS consists of two methodological steps that have analogs in a GWAS”, where “environmental assays” are also called environmental “loci” (Patel et al. 2010, p. 2).4 For now, we can conclude that the literature about the exposome is intended to better highlight of the role of the environment in the emergence of non-communicable diseases. This objective, however, is paradoxically at the origin of imprecision and difficulties regarding the very concept of environment. This constitutes the topic of the next section.

3 Conceptual Shortcomings of the Environment as Exposome The definition of the environment as exposome (i.e., as anything that is non-genetic) raises a number of conceptual issues highlighted and discussed in this section. We show that this understanding transforms the concept of environment into a catch-all and, consequently, a nonoperational category. We also underline that it leads to the wrong scale of analysis, insofar as it tends to refer to that which surrounds genes, thereby characterizing the environment of the genes rather than the environment of the entities susceptible to being healthy or sick (i.e., human individual organisms or populations).

4 For an in-depth analysis of the pitfalls of the symmetry between the exposome and the genome, see Giroux (2021a).

FROM EXPOSOME TO PATHOGENIC NICHE. LOOKING …

3.1

181

Environment in Exposomics: A Wide and All-Encompassing Concept

Beyond the theoretical reasons outlined above, the broad definition of the environment as anything non-genetic is likely to be linked to the ambition in exposomics to study the totality of individuals’ exposures (Siroux et al. 2016). In other words, it may be related to the holistic dimension of the exposome (Vrijheid, 2014, p. 877; Johnson et al. 2017). However, such characterization implies that the environment includes an innumerable amount of diverse elements, such as non-genetic parts of the body, as well as all kinds of immediate and remote surroundings, both in space and time, which significantly expands its spatio-temporal limits (for a similar analysis, see Guchet (2019, p. 75). Not surprisingly, various papers offer cumulative lists of factors included in the exposome (and therefore in the environment). They mention elements such as “air quality, industrial chemicals, pesticides, radiations, food and water, tobacco, medication, physical activity, climate, psychosocial stress, urban environment, social support” (Niedzwiecki et al. 2019, p. 115), refer to “exposures from the environment, diet, behavior, and endogenous processes” (Miller and Jones 2014, p. 2), evoke “where we live, what we eat, drink, or breathe, our social economic status, behavior, social interactions, occupation, and exposure to pollutants” (Envirome) (Toscano and Oehlke 2005, p. 12), and talk about “the set of air pollutants, noise, meteorological factors, green spaces, and built environment”5 (Robinson et al. 2018, pp. 1–2), and mention how the EWAS can target up to 266 factors (Patel et al. 2010). Some authors even refer to how humans experience more than a million exposures in a lifetime (Johnson et al. 2017). They also stress how a “detailed understanding of life-stage-specific effects represents an enormous challenge” (ibid.). However, one can question the extent to which such an exhaustive ambition is epistemologically and pragmatically suitable (see Giroux 2021a). First, from a methodological point of view, an exhaustive measurement of environmental factors, and in particular the integration of all these measurements, appears technically complicated, if not impossible (Wild 2005, p. 1848; Siroux et al. 2016; Johnson et al.

5 Here the authors are interested in the Urban exposome, defined by 28 factors.

182

G. PONTAROTTI AND F. MERLIN

2017).6 Second, and more importantly, a boundless and all-encompassing concept of environment—based on a strictly cumulative approach devoid of general conceptual reflection—might not be useful either for better accounting for the origin of diseases or for controlling health problems. For example, in the case of obesity, “attempting to consider every possible environmental contribution to energy balance can quickly become overwhelming” (Kirk et al. 2010, p. 116). This expresses how explanatory relevance and the interventional effectiveness it can provide do not derive from precision in the sense of an exhaustive listing or measurement of the factors involved in the production of a given phenomenon. On the contrary, it suggests that they might rather be based on discriminating between causal factors that are more or less strongly involved in the production of the phenomenon. Let us clarify our argument and apply it to human health studies. First, we argue that in human health studies, a relevant explanation does not have to be precise in the sense of offering an exhaustive characterization of the environmental elements involved in the production of an event. Rather, it should be efficient, in the sense that it must open the way to robust predictions and possible interventions.7 Indeed, the epistemic objective of this kind of scientific research, which is to account for the etiology of diseases, is directly linked with their pragmatic objective, which is the control of health problems via interventions grounded in social and political concerns. More generally, scientists interested in human health are supposedly willing to identify causes appearing as variables that can be manipulated in order to modify their effects (in this case, the health status of human populations) and might therefore endorse an interventionist account of causation (see Woodward, 2003; Pearl 2009).8 Second, philosophy of science, and more particularly the debate about the distinction between causes and conditions (Mackie 1965; Cheng and

6 Note that this first shortcoming is explicitly assumed in the literature on the exposome, in particular by Wild (2012) who admits that the exposome possibly taken into account will always be partial. 7 The relevance of an explanation can be grounded on other elements (e.g., simplicity, integration into a larger theoretical network, etc.). 8 According to these accounts: “causal relationships are relationships that are potentially exploitable for purposes of manipulation and control: very roughly, if C is genuinely a cause of E, then if I can manipulate C in the right way, this should be a way of manipulating or changing E” (Woodward, 2016).

FROM EXPOSOME TO PATHOGENIC NICHE. LOOKING …

183

Novick 1991), reminds us that identifying the causes of a disease (be they environmental or other) differs from identifying the sum of the conditions that are needed for its occurrence. This debate more precisely shows the epistemological necessity for discriminating among a set of facilitating conditions when trying to build relevant (sensu efficient) scientific explanations. In this view, a concept of environment that includes all non-genetic causal factors does not appear suitable for use in relevant and efficient causal explanations—namely in explanations that, in focusing on key determinants, allow efficient intervention and control of phenomena. Finally, while the literature on exposomics tries to better highlight the role of the environment in the etiology of some diseases, it paradoxically outlines a concept of environment that turns out to be very broad and not operational insofar as, by including everything, it impedes efficient focus on key determinants in theoretical explanation and practical interventions.9 3.2

The Disappearance of the Organism and the Environment of the Genome

In the exposomic literature, the fact that the environment as exposome is equated with everything that is non-genetic has two other related and unwelcome consequences. Firstly, equating the environment with non-genetic factors surprisingly induces the subtle disappearance (or eclipse) of the organism from discourse around human health. More explicitly, reducing the organism to

9 As noted by one of our reviewers, biological events are often determined by a vast

array of small causes, including when these causes are genetic, and not necessarily by some key determinants (as reminded by the literature about missing heritability). However, we maintain that for practical reasons, it remains necessary to discriminate among the environmental factors involved in the etiology of a disease, at least to set a priority order of interventions.

184

G. PONTAROTTI AND F. MERLIN

its genome10 (which is one of the outcomes of omics biomedicine)11 leads to a pair of classical interactants, composed of organism and environment (O/E) (Lewontin 2001; Pearce 2014), to be replaced with a different pair, of genome and environment (G/E). Symptomatically, Toscano and Oehlke (2005, p. 12) propose using the term “enviromics” to designate the “interactions of the complete environment, or envirome, with human genomes”. In doing so, they suggest that the entity that interacts with its environment is not the human individual (organism), but rather the human genome. The replacement of one pair of interactants with another appears even more clearly in other contributions. For example, when arguing that “(a)ssessing environmental exposures with a similar degree of accuracy would enable efficient study of the interplay between the environment and the genome”, Siroux and colleagues (2016, p. 125) clearly state that the relevant interactions to be taken into account involve genomes and environments. Finally, the theoretical disappearance of the organism is evident in the exposomic interpretation of nurture in which the exposome “captures the essence of nurture” and is defined as “the summation and integration of external forces acting upon our genome throughout our lifespan” (Miller and Jones 2014, p. 2). Second, and relatedly, use of the concepts of “internal exposome” and “internal environment” (see Sect. 2)12 corroborate how the relevant environment of exposomics is actually that of the genome (and the related disappearance of the organism). To make sense of these concepts, which initially appear as counterintuitive insofar as the environment is usually expected to be outside the body (Last 1995), one option is to consider 10 According to the reductionist approach in exposomics, the individual organism and the environment can be characterized and measured by elements found in blood samples, namely DNA sequences and biomarkers. The idea here is that molecular biomarkers can traduce or reflect, or even constitute the environment, which is actually that of the genome. 11 Personalized medicine is often reproached for this sort of reduction. Indeed, as shown by Guchet (2019), this medicine is grounded on the assumption that the genomic characteristics of the patients determine the kind of treatment they should take. It is thus clearly directed towards the genomes of individuals. According to Guchet, exposomics participates to the elaboration of precision or personalized medicine when it characterizes individual environmental exposure at the individual level by using omics technologies (such as transcriptomics, proteomics, metabolomics, etc.). 12 For example, Wild’s internal exposome (2012) which includes metabolism, endogenous hormones, body morphology, but also gut microflora inflammation clearly refers to the environment of genes, just as Rappaport’s internal environment (2011).

FROM EXPOSOME TO PATHOGENIC NICHE. LOOKING …

185

the internal environment as merely a “reflection” of external exposure.13 In this view, the internal environment should not be conceived as an environment per se, but rather as a translation or reduction, or more simply as a molecular manifestation of environmental exposure, which is more easily measurable by biomarkers inside the body (see Siroux et al. 2016, p. 125). Another option however consists of admitting that the environment outlined in exposomics is not that of organisms but that of the genome, given that, in exposomic literature, gene regulation is what matters for making sense of the development of diseases. Our objection to this view is twofold. First, the question of health and disease can only be addressed for human individuals (i.e., organisms) or human populations (Rose 1985). Second, the environments of genes, of organisms and populations of organisms do not have the same spatio-temporal boundaries, and should therefore not be conflated. This being said, none of these environments are independent—they reciprocally influence each other. These are the reasons why, in what follows, we propose conceiving of the environment relevant for health studies in a more dynamic and relational way, and at the right scale of analysis. By using the tools developed in philosophy of biology, we will show in Sect. 4 that the relevant environment in health studies is that of a population, and that the limits of this environment depend on its dynamic, dialectical relationship with this population.

4

Conceiving the Pathogenic Environment with the Tools of Philosophy of Biology

Philosophy of biology, we argue, can provide tools to shape a relevant concept of the environment in general, and of the pathogenic environment in particular. Indeed, philosophical reflections on this topic highlight that the environment is a relational concept, that the relevant environment for living beings may be thought of in terms of constructed niche, and that this constructed niche can be conceived at various scales, both spatial and temporal. In what follows we show that the pathogenic environment can—and should—also be thought of in these terms, as a pathogenic constructed niche. 13 For instance, Giroux (2021a, p. 131) argues that this represents a form of methodological reductionism where one of the parts (the internal environment) is taken for the whole (the exposome or the environment in all its dimensions).

186

G. PONTAROTTI AND F. MERLIN

4.1

The Environment in Philosophy of Biology

The concept of biological environment is relational, in both a weak and a strong sense. In a weak sense, the concept of biological environment is relational insofar as this environment is always defined with respect to a reference entity. Indeed, it designates what surrounds a focal entity, be it a gene, an organism, a population of organisms, a species, etc. As noted by biologist R. Lewontin (2001, p. 48), “An environment is something that surrounds or encircles, but for there to be a surrounding there must be something at the center to be surrounded”. In a stronger sense, the concept of environment is relational because the environment is an element that is always more less actively determined by the encircled biological entities (Lewontin 1983, 2001; Odling-Smee et al. 2003). In this respect, the environment of an organism is not stricto sensu what surrounds this organism, namely the sum of external conditions in which it has to make a living, but rather “the penumbra of external conditions that are relevant to it because it has effective interactions with those aspects of the outer world” (our emphasis) (Lewontin 2001, pp. 48–49). The relevant surrounding of an organism, such as an animal, is first determined by “the properties of the animal’s sense organs, nervous system, metabolism and shape” (p. 52). It is therefore “the space defined by the activities of the organism itself” (p. 53). The example provided by Lewontin is that of the environment of three bird’s species (pp. 51–52). Even if these species share the same geographical surrounding, they do not have the same biological environment: In my garden there are trees, and grass growing around the trees, and some stones lying here and there on the ground. The grass is part of the environment of a phoebe, a bird that makes its nest out of dried grass, but the stones are not part of its environment. If they disappeared it would make not the slightest difference to the phoebe. But those stones are part of the environment of a thrush, a bird that uses the stone as an anvil to break open snails on which it feeds. There are holes high up in the tree which woodpeckers use for nests, but these holes are not part of the environment of either the phoebe or the thrush. The elements of each bird’s environment are determined by the life activities of each species. (p. 53)

Besides, the relevant external conditions of an organism can be constructed by the organism itself through its activities (p. 54). This

FROM EXPOSOME TO PATHOGENIC NICHE. LOOKING …

187

phenomenon is generally referred to as “niche construction”14 and has been the subject of abundant literature (Odling-Smee et al. 2003; OdlingSmee 2010). Niche construction more precisely refers to the fact that an organism modifies the factors constituting its environment “either by physically perturbing these factors at its current address or by relocating to a different address, thereby exposing itself to different factors” (Laland et al. 2001, p. 118). Nests, dams and burrows are key examples of constructed environments. Finally, organisms modify their environment in the sense that they transform what they eat into waste, which is a resource for other species. In this view, “living systems are the transformers of materials, taking in matter and energy in one form and passing it out in another that will be a resource for consumption for another species” (Lewontin 2001, p. 55). To summarize, there is a dialectical relationship between organisms and their environment insofar as these two elements can be said to shape each other (see Lewontin (1983) for a complement). Biological environment is not an equivalent of physical or geographical environment, as noted by Uexküll (1965).15 It should not be considered as a separate and independent entity, but rather as a constructed niche that is determined by the metabolism, senses, shape and activities of organisms (or any reference entity which can be considered as an agent). In brief, the environment is always ontologically dependent on its reference entity. Constructed niches can be conceived at different scales. One can distinguish locally designed surroundings, which are actively maintained by individual organisms (or groups of organisms), and globally transformed surroundings, which are collectively produced by a huge number of organisms (Sterelny 2005). In the first case, the effect per capita is important; in the second case, it is small. An example of an individually designed environment is beaver dams, while an example of a collectively transformed environment is the oxygen atmosphere (or forest soil, which is modified by microorganisms).

14 The concept of niche comes from ecology, the science that studies the interactions of organisms with their surroundings. It has been defined as a property of an ecosystem or a set of variables that allows a species to make a living. The concept of niche has further been developed in evolutionary biology, where specialists have conceived about niche construction (for a more detailed appraisal, see Pocheville 2015). 15 The ethologist distinguishes the geographical world, the Umgebung, and the world defined by actions and perceptions of organisms, namely the Umwelt (for a general analysis, see Canguilhem 2003).

188

G. PONTAROTTI AND F. MERLIN

The concept of biological environment also depends, from an epistemological point of view, on the specific problems addressed by scientists. In other words, the concept refers to different elements and takes on different meanings according to the questions addressed by biologists (Pontarotti et al. 2022). For example, in evolutionary biology, which is framed by the theory of evolution through natural selection, the environment is composed of a set of selective pressures. In other words, “(f)rom the point of view of the theory of natural selection, the relevant environment is the selective environment” (Brandon 1990, p. 84). The selective environment is distinguished from the “external environment” which refers to all factors external to the population of interest (p. 83). It specifically “comprises the complete set of external conditions that are held constant across members of a population at a time, and are causally relevant to the fitnesses of types” (Walsh 2022, referring to Brandon’s 1990 definition). In this view, the environment should be common to different individuals for there to be differential reproductive success among them (see also Antonovics et al. 1988). Actually, the very concept of constructed niche can refer either to selective niche or to developmental niche (Stotz 2017). The selective niche refers to the fact that organisms modify the selection pressure acting on them and on their offspring through their environment-changing activities (OdlingSmee 2010). The developmental niche is considered in reference to the ontogenetic niche (West and King 1987). It designates “the non-genetic (exogenetic) yet heritable factors influencing an organism’s development” and contains “reliably but flexibly inherited physical, social, ecological and epistemic resources needed to reconstruct, or modify, that developmental system” (Stotz 2017, p. 3). The selective niche is relevant for evolutionary biologists, the ontogenetic niche is a useful theoretical tool for scientists interested in ontogenesis and/or considering ontogenesis as a key process for evolution (Stotz 2017; Bateson 2005; Laland et al. 2015). In any case, the ontogenetic niche and the selective niche do not have the same definition. Consequently, they do not comprise the same elements, nor do they necessarily overlap (Stotz 2017). To summarize, studies developed in philosophy of biology underline that the environment is not merely what surrounds a reference entity, be it a genome, an organism or a population of organisms. As a result, it cannot be conceived as everything that is not this entity. More specifically, these studies provide the following double lesson: (1) the environment

FROM EXPOSOME TO PATHOGENIC NICHE. LOOKING …

189

is always more or less actively determined by its reference entity, therefore appearing as a constructed niche (ontological dependence); (2) this niche comprises different elements according to the problem addressed by scientists (epistemological dependence). We argue that the concept of environment in human health studies, if it is intended to be epistemologically relevant and pragmatically useful, should be shaped with this double lesson in mind. More than that, this double lesson would empower the concept of environment, for example by avoiding the shortcomings of environment conceived as exposome (see Sect. 3), and by compelling scientists to focus on key determinants in their models and intervention plans. Besides, because the environment in question is a constructed niche, it de facto suggests that the reference entity could have the power to modify it. 4.2

Towards a Concept of “Pathogenic Niche”?

Let us apply the double lesson from philosophy of biology in order to define a notion of environment that is relevant and operational in exposomics and, more generally, in human health studies. The result is the following: first, the concept of pathogenic environment is, ontologically speaking, relational with respect to the focal entity it surrounds, which is a human organism or a human population. Bearing in mind that exposomics takes on board the concerns of both toxicology and epidemiology, and thus also aims to contribute to public health studies (cf. Introduction, see Wild 2005), we decide to consider a human population as the focal entity surrounded by a given environment.16 Each human population, by its properties and activities (which are genetic, behavioral, socio-historical, economic, political, cultural, etc.),17 determines

16 Rose (1985) distinguishes what he calls the question of “the causes of cases” at the individual level and the question of “the causes of incidence” at the populational level. Even though epidemiology is commonly defined as the study of the determinants of the distribution of the disease, Rose recalls that epidemiologists come to address both questions. On the basis of his analysis of the advantages and disadvantages of the individual and of the population strategy, Rose shows that the latter “has a large potential—often larger than one would have expected—for the population as a whole” (p. 37). He thus argues that, from the viewpoint of public health, “the priority of concern should always be the discovery and control of the causes of incidence” (p. 38). 17 In line with ecological approaches in epidemiology (e.g., Susser & Susser 1996), by properties of (human) population we do not mean mere statistical aggregates of individual

190

G. PONTAROTTI AND F. MERLIN

and builds its own pathogenic environment. Moreover, it alters it, potentially having an impact on the pathogenic environment of other human populations and/or of other species. Second, the concept of pathogenic environment is relational, epistemologically speaking, with respect to the interests and objectives of researchers working on human health who address the question of the etiology of human diseases by particularly focusing on the environmental determinants of non-communicable ones. Thus, the pathogenic environment of a human population corresponds to the subset of factors surrounding the defined population, that are determined/built/modified by it (ontological dependence),18 and that participate in the explanation of the origin of diseases affecting it (epistemological dependence). Let us be more specific by considering the example of an environmental factor which is potentially pathogenic for some human populations but not others (because of these populations’ properties). Namely, a factor which is part of the pathogenic niche of one population and not of another, even if these two populations globally share a geographical location. Let us consider the sun, which is to say the UV rays that all human populations are surrounded by but exposed to in greater or lesser degrees depending on their respective geographical location. Are UV rays part of the pathogenic environment of any human population? The answer to this question lies in both the ontological and epistemological dependence of the concept of environment. As for epistemological dependence, if the question we are interested in is, for instance, the pathology of skin cancer, UV rays should be considered a good candidate for being part of

properties, but characteristics of the population itself as a whole that are neither measurable nor reducible to the level of individual organisms (i.e., “integrated variables”, such as population density, income distribution, the distribution of socio-economic inequalities, social norms, vs. “derived or aggregated variables” directly inferred from observations at the individual level, such as mean number of smokers and mortality rate). According to ecological approaches, populations are thus the relevant scale of analysis and inference, hence the relevance of the concept of population health. See Giroux 2008, 2021c. See also Rose 1985. 18 Note that, in the niche construction literature, what we prefer to gather together under the expression “ontological dependence” (i.e., the fact that the organism and the environment are co-defined with the physical niche construction) can be presented as distinct and respectively considered as the layers of ontology (ontological determination) and of causation (causal modification), the layer of epistemology corresponding to what we call “epistemological dependence” (e.g., see Godfrey-Smith, 1998). We thank one of the reviewers for this remark.

FROM EXPOSOME TO PATHOGENIC NICHE. LOOKING …

191

the pathogenic environment of a given human population, given the data we already have about the relation between sun exposure and skin cancer. It goes without saying that, if the disease being studied is unlikely to be affected by UV rays (an example of which, we admit, is difficult to find), they should not be considered as a pathogenic environmental factor. What about ontological dependence? What about the relationship between UV rays and human populations in the context of epidemiological research on skin cancer? The way each population (1) determines, (2) builds and (3) modifies its own environment provides an answer to this question. First, UV rays can be part of the pathogenic environment of a population (or not) depending on its properties, for instance its genetic and hormonal features at the origin of melanin levels or its socio-professional category. Let us look closer at this second example by considering a population of miners and a population of farmers living in the same area (e.g., the city of Meekatharra in West Australia). The two populations are differently exposed to the sun because of their respective professional activities. Miners work during the day inside the mine and thus are not exposed to sun (but are exposed to other pathogenic factors), while farmers, who work most of the time in the daylight, are highly exposed to UV rays. We can thus conclude that, with respect to skin cancer, the sun is part of the pathogenic environment of farmers but not of miners.19 Note however that such a conclusion is a matter of degree, the population of miners is exposed to UV rays too, but to a far lesser extent than the population of farmers. Second, UV rays are pathogenic environmental factors in a given population depending on the activities the population performs in order to build its own (physical, political) environment. For instance, imagine two populations living in the same geographical area characterized by high UV rays’ intensity (Perth in Western Australia). They live in the same city but in different neighborhoods. Although again it is a matter of degree, the population living in a neighborhood where trees have been planted and sun shelters have been built no longer necessarily counts UV rays as part of its pathogenic environment. In fact, because of planning policies

19 As interestingly suggested by one of the referees, low exposure to UV rays could nonetheless be part of miners’ pathogenic environment with respect to other diseases (other common cancers, autoimmune diseases, infectious diseases, and cardiovascular diseases) because of the vitamin D deficiency it provokes (see Holick 2020).

192

G. PONTAROTTI AND F. MERLIN

and/or votes cast by the population20 —in other words, because of this population’s socio-political activities—sun exposure has been limited, its impact in terms of skin cancer incidence in the population has been too. This is not the case for the population living in a different neighborhood in the same city, where no such activities have been performed. In this case, UV rays are likely to be part of the pathogenic environment of the population with respect to skin cancer, which continues to have a high incidence in the population because of excessive sun exposure. The same can be said if we compare, again, two populations living in the same city but different neighborhoods, one in which an information and prevention campaign on the risks of sun exposure has been carried out, the other not. Sun exposure does not have the same chance of actually being part of the pathogenic environment of the two populations because of the different governance decisions made by each neighborhood’s population. Third, the question of whether UV rays are part of the pathogenic environment of a given population partly finds its answer in this population’s and/or other human populations’ activities, for instance manufacturing activities that have chemically altered the atmosphere, producing ozone depletion.21 Indeed, one of the consequences of the phenomenon of depleted ozone levels is an increase of UV radiation, especially for populations living in mid-latitudes areas of the planet. For example, based on a recent report by the Ecuadorian Space Agency, UV radiation is likely to be part of the pathogenic niche of Ecuadorian human populations with respect to skin cancer.22 We thus argue that, in order to be more specifically and operationally defined, the pathogenic environment should be conceptualized as a niche, i.e., the pathogenic niche. This niche corresponds to the set of elements that, at various micro- and macro-levels and throughout time, surround a population, that are determined, built and altered by this population’s

20 For an example of policies aimed at preventing high levels of sun exposure in Australia, see the second edition of “Creating Shade at Public Facilities - Policy and Guidelines for Local Government” prepared by the School of Public Health, Queensland University of Technology and funded by Queensland Health in Australia in 2012 (https://www.health.qld.gov.au/__data/assets/pdf_file/0020/422264/20267.pdf). 21 As noticed by one of the reviewers, this third example of human populations’ activities altering the environment targets a very different and higher scale than the previous one about the use of tree shade to prevent too high levels of sun exposure. 22 http://exa.ec/bp21/index-en.html. See also Parra et al (2019).

FROM EXPOSOME TO PATHOGENIC NICHE. LOOKING …

193

properties (genetic, metabolic, behavioral, social, economic, political, cultural), and that account for a high disease incidence in that population. The pathogenic niche is thus a relational and non-static entity, dynamically defined by the dialectical relationship it has with the reference population over time. Moreover, it includes non-genetic elements but does not equate to everything that is non-genetic, and can also include genetic elements (e.g., a pathogen’s genome). Its spatio-temporal boundaries are thus more restricted than those of the exposome. Finally, it spans multiple scales, from the molecular to the socio-cultural, and takes into account their interactions, without reducing them to each other. According to this view, the environment that matters for health sciences is always built at different scales, which should be the focus of relevant interventions. The concept of pathogenic niche, we argue, is more operational than the concept of exposome, both from a theoretical and a pragmatic point of view, for two main reasons. First, thanks to its restricted spatio-temporal boundaries, the concept of pathogenic niche would be a constant guide for scientists in the elaboration and refinement of epidemiological studies, one which would finally allow them to focus on key health and disease determinants in their models and intervention plans. This statement is grounded in the view that including too many factors impoverishes scientific explanations (Giroux 2021a) and, consequently, practical application, as well as the view that explanatory relevance and pragmatic effectiveness do not come from exhaustiveness but rather from discrimination23 (see Sect. 3.1 above). Second, because the concept of pathogenic niche implies that the environment is (at least partly) produced by humans, it follows that it can be controlled and modified provided that the right scale of analysis and intervention is identified. From this perspective, the concept of environment offers a more operational as well as a more dynamic and empowered (but also empowering) concept to health scientists.

23 Discrimination based on ontological and epistemological dependence in the case of the pathogenic niche.

194

4.3

G. PONTAROTTI AND F. MERLIN

The Concept of Pathogenic Niche in the Wake of Ecological Approaches in Public Health and in Epidemiology

While the pathogenic niche approach maintains distance from the results of exposomics regarding the concept of environment (i.e., with the equivalence between environmental and non-genetic), it somehow appears reminiscent of ecological approaches in health sciences on two counts: (1) It echoes socio-ecological models developed in public health (McLeroy et al. 1988; Stokols 1992; Stokols et al. 1996), and (2) it partially meets ecological approaches developed in epidemiology, namely ecoepidemiology (Susser and Susser 1996; March and Susser 2006; Bizouarn 2016) and ecosocial theory (Krieger 1994, 2001). We briefly expand on these two counts below. (1) In the field of public health (and behavioral) sciences, ecological models “emphasize the environmental and policy contexts of behavior, while incorporating social and psychological influences”. Put another way, these models “lead to the explicit consideration of multiple levels of influence” on health behavior, “thereby guiding the development of more comprehensive interventions” (Sallis et al. 2008, p. 465). Ecological models are willing to “focus on the nature of people’s transactions with their physical and sociocultural surroundings, that is, environments” (p. 266). However, they are mainly based on the assumption that complex, built and multi-scale environments determine individual behavior, and not vice versa, as shown by the recapitulative table proposed by Sallis and colleagues (2008, pp. 468–469). For example, the ecological model for health promotion proposed by McLeroy and colleagues (1988) and included in Sallis et al.’s table, is based on a multilevel analysis of health determination (interpersonal, organizational, community, public policy), which should lead to a multi-scale strategy of intervention. However, it does not explicitly highlight the fact that population and environment co-determine each other. By contrast, and quite originally, the social ecological paradigm for health promotion developed in public health studies by Stokols and colleagues (1992, 1996) explicitly takes into account the codetermination of population and health environments. This paradigm proposes a global and multi-scale approach to health promotion and aims at “understanding the complex community and environmental origins of public health problems” (Stokols et al. 1996, p. 247). Borrowing from ecology’s core hypothesis of an interrelation between organisms

FROM EXPOSOME TO PATHOGENIC NICHE. LOOKING …

195

and their environment, the paradigm is grounded on the following key assumptions (Stokols 1992, pp. 7–8). Health promotion should rest on a more global analysis that encompasses consideration of the interplay between different factors (biological and notably genetic, physical, natural and technological environment, as well as social and political environment); it should also integrate the complex and multi-scale dimension of human environments and acknowledge various levels of participants in the environment (individuals, groups, organizations, etc.). Finally, and importantly, health promotion must take into account the interactions and mutual determination between the people and their environment. Thus, people-environment transactions are characterized by cycles of mutual influence, whereby the physical and social features of settings directly influence their occupants’ health and, concurrently, the participants in settings modify the healthfulness of their surroundings through their individual and collective actions. (Stokols 1992, p. 8, our emphasis)

(2) In the field of epidemiology, the theoretical framework of ecoepidemiology (Susser and Susser 1996; March and Susser 2006; Bizouarn 2016) aims to make the discipline relevant for public health issues. The eco-epidemiology framework was proposed as an alternative to the risk factor approach, and to be in line with the populational epidemiology of the 1950’s (see Morris 1955). Contrary to risk factor epidemiology, which focuses on individuals, eco-epidemiology aims to integrate various levels of causation (molecular, social, etc.) and their relations. It relies on the metaphor of Chinese boxes, in which each box represents a level of causation which is embedded in larger ones and includes smaller ones. In social epidemiology, the intention of ecosocial theory is to integrate various levels of causation and, notably, to articulate biological and social determinants of disease (Krieger 1994, 2001). Ecosocial theory rests on a mixed metaphor—the scaffolding of society and the bush of evolution—that opens a way to conceptualizing “epidemiologic profiles” linked with social structure, cultural norms, ecologic milieu and genetic variability (1994, p. 897). Finally, ecosocial theory integrates individual and population levels. According to Krieger, it “requires population-thinking in its study of individuals, and recognition of individual variability (and similarity) in its study of populations”. Eco-epidemiology and ecosocial theory are associated with “mental pictures” that are “both multidimensional and dynamic”, as well as with a terminology that invokes “notions

196

G. PONTAROTTI AND F. MERLIN

of ecology, situating humans as one notable species among many cohabiting, evolving on, and altering our dynamic planet” (Krieger 2001, p. 671). The concept of pathogenic niche clearly echoes all of these ecological perspectives in health studies. Like them, it is grounded on the core hypothesis of ecology, according to which organisms actively interact with their surroundings. Besides, it embraces a multilevel and heterogeneous conception of the environment involved in health and disease, and endorses a processual perspective, in which the environment that matters for health sciences is dynamic. Finally, it takes into account the population scale. However, as shown above, most ecological approaches insist on the fact that health is determined by a set of environmental factors, and seem to leave in the theoretical background the fact that the environment affecting the health of the population is determined by this very population, by virtue of dialectical movement. In other words, they underline that the health of a population is co-determined by various environmental factors (joint causality), but they do not explicitly stress the fact that environment and population mutually determine each other (co-construction), except for Stokols and colleagues’ social ecological approach. As a result, the concept of pathogenic niche appears more specifically in accordance with the latter. In brief, most ecological models do not fully and explicitly admit the lesson of ontological dependence provided by philosophy of biology. Their lack of explicit consideration for co-construction appears as surprising insofar as they emerged in the context of an awareness of humans’ impact on the environment (Kickbusch 1989). However, this can be explained by the fact that ecological approaches in health studies mainly aim to better understand the role of the environment in the determination of disease in a population rather than investigate the way in which the population shapes its surroundings. 4.4

The Obesogenic Environment as a Pathogenic Niche

Before concluding, we will now examine the way the environment is conceived in the literature about obesity. This example provides a very good case, we hope, for underlining the relevance and usefulness of the concept of pathogenic niche.

FROM EXPOSOME TO PATHOGENIC NICHE. LOOKING …

197

Obesity is usually defined as a disease characterized by an imbalance between energy intake and energy expenditure (Vrijheid et al. 2020). Some authors have argued for the need to better take into account the role of the environment, specified as “obesogenic”, in the development of this disease in individuals and populations (Swinburn et al. 1999; Kirk et al. 2010). These authors have lamented “the lack of suitable paradigms and tools for understanding and measuring the environment” (Swinburn et al. 1999, p. 564). They have, more precisely, regretted that “(d)espite the explosion of obesogenic environment research within the last decade, consensus on what constitutes the very environment we are trying to measure has not yet been reached” (Kirk et al. 2010, p. 109). Besides, they have underlined the difficulty of offering a clear definition of the obesogenic environment insofar as “The environment is a nebulous concept, being complex, dynamic and multi-level” (Kirk et al. 2010, p. 110). Finally, they have noted that the vagueness of the concept of environment makes it difficult to identify relevant opportunities and a scale of action. In this view, a relevant concept of environment should be precise and discriminatory (vs. all inclusive) given that “attempting to consider every possible environmental contribution to energy balance can quickly become overwhelming” (Kirk et al. 2010, p. 116). The obesogenic environment has explicitly been defined as “the sum of influences that the surroundings, opportunities, or conditions of life have on promoting obesity in individuals or populations” (Swinburn et al. 1999, p. 564). It has been conceived of at micro and macro scales and divided into different categories, notably in the context of the ANGELO framework (analysis grid for environments linked to obesity). This framework, which is intended as a “conceptual model” and “practical tool” for relevant interventions (Swinburn et al. 1999, p. 563), is described as a “2 × 4 grid which dissects the environment into environmental size (micro and macro) by type: physical (what is available), economic (what are the costs), political (what are the ‘rules’), and sociocultural (what are the attitudes and beliefs)” (p. 563). The ANGELO framework was built in line with an ecological approach to the obesity pandemic (Egger and Swinburn 1997). Based on the assumption that educational interventions are not sufficient for reducing the incidence of obesity in a population, this ecological approach outlines “three main influences on equilibrium levels of body fat—biological, behavioural, and environmental” (1997, p. 477). The obesogenic environment has also been more implicitly characterized as an environment that “promotes high energy intake and sedentary

198

G. PONTAROTTI AND F. MERLIN

behaviour” (Lake and Townshend 2006, p. 262). The authors who have proposed such characterization distinguish among the “built environment” (physical design, land-use patterns and transportation systems) and the “food environment” (availability, advertising, “food prepared and consumed at home, and out of home sources”, “vending machines, takeaways, cafes, restaurants, supermarkets and convenience stores”). Other studies dedicated to the obesogenic environment have focused on various environmental factors, notably “access to food sources, access to recreational places, opportunities for utilitarian physical activity, perceived safety, and neighborhood sociodemographics” (Wall et al., 2012) but also “greenspace, walkability, supermarket density, unhealthy food outlet relative density, spaces for social interaction and air quality” (Wilding et al., 2020).24 Beyond the explicit references, the literature on obesogenic environment seems to be clearly influenced by ecological models of health promotion. First, it insists on the fact that various and multi-scale environmental elements co-determine the incidence of obesity in the population. Notably, it focuses on “the influences of the built environment on physical activity and dietary behaviours” (Townshend and Lake 2017, p. 38) and mentions the need to implement a “collaborative strategy with the multiple sectors which impact on the problem” (Egger and Swinburn 1997, p. 477). This literature also describes an environment that is built by human populations, even if it does not make explicit that the population determines its environment and mainly highlights the influence of environment on behavior. For example, the ecological model proposed by Egger and Swinburn (1997) outlines three kinds of built environment that influence food intake and physical activity: physical environment (e.g., food technology, fitness industry policies, proximity of fast-food outlets, etc.), economic environment (e.g., food taxes, family incomes, gym or club fees, etc.) and socio-cultural environment (e.g., consumer demand, family eating patterns, school, attitude to sport, etc.). The

24 It is also worth mentioning more general studies adopting the approach called Geometric Framework for Nutrition (GFN). The aim here is to interpret the way that nutrients influence physiology and health (Simpson et al, 2017). In particular, GFN models take into account one or more foods, the organism’s current and optimal nutritional states, the impact on its nutritional state of eating each food, its body composition, the efficiency of nutrient utilization, the rate of excretion and whatever performance consequences might be of interest (cf. Raubenheimer et al., 2009).

FROM EXPOSOME TO PATHOGENIC NICHE. LOOKING …

199

ANGELO framework, which is based on this ecological approach, recognizes that the micro-environmental settings (schools, workplaces, etc.) are “influenced by individuals” while “(m)acroenvironmental sectors” are “groups of industries, services or supporting infrastructure that may influence diet, physical activity or weight” (Swiburn et al. 1999, p. 565; Kirk et al. 2010, p. 110). We argue that the obesogenic environment should be qualified as an “obesogenic niche”. Designating it as such would not constitute a mere re-labeling. It would have the advantage of making explicit that an obesogenic environment is fundamentally a dynamic entity constructed by a set of agents at several levels (groups with specific genetic susceptibility, social groups, political institutions and private companies, etc.). In stressing the dialectical relation—namely co-determination—between population and environment, the concept of obesogenic niche would make the notion of environment more explanatorily relevant and more operational, pragmatically speaking, by offering a way to think about efficient interventions at different scales. It would provide a conceptual tool that better allows for agencies and responsibilities to be pointed out. Finally, it would offer a consistent framework to integrate different parts of this built environment and make sense of their relationships. For example, the concept of an “obesogenic niche” could relate food with built environment by showing how they depend on each other.

5

Conclusion

In this chapter we first showed the main shortcomings of the concept of environment as exposome as too wide and all-encompassing a concept, one which is defined at the level of individual genomes and provokes the disappearance of the organism. We then argued for the consequent need to adopt a less broad but more relational and dynamic perspective to reconceive the spatio-temporal limits of the environment in health studies. More precisely, we argued that, in order for the concept of environment to be more conceptually and operationally relevant in this research context, it should be conceived as a pathogenic niche, epistemologically shaped by the physicians’ questions, and ontologically determined by its interactions with the population that inhabits it. Grounded on lessons drawn from the literature in philosophy of biology about how to conceive of and define the environment, the concept of pathogenic niche not only fuels the philosophical debate about niche construction, it can also

200

G. PONTAROTTI AND F. MERLIN

allow for it to expand into the domain of health studies (in particular exposomics and, more broadly, epidemiology and public health) where ecological perspectives already exist, but seldom foreground the idea of a mutual, dialectical determination between human populations and their environments. Last but not least, the concept of pathogenic niche is also intended as a tool for health scientists to scaffold hypotheses about which environments should be studied and considered in order to plan efficient health interventions. Acknowledgements We would like to thank for his contributions and feedbacks one anonymous reviewer, and our collaborator Élodie Giroux for her fruitful comments. This work was supported by the ANR EnviroBioSoc Project (grant 19-CES26-0018-01 of the French National Research Agency).

References Antonovics, J., Ellstrand, N. C., & Brandon, R. N. (1988). Genetic variation and environmental variation: Expectations and experiments. In: Plant evolutionary biology (pp. 275–303). Springer, Dordrecht. Baccarelli, A. A. (2019). The human exposome: A new “omic” ready for prime time. Journal of the American College of Cardiology, 74(10), 1329-1331. https://doi.org/10.1016/j.jacc.2019.07.029 . Bateson, P. (2005). The return of the whole organism. Journal of Biosciences, 30(1), 31-39. https://doi.org/10.1007/BF02705148 . Bizouarn, P. (2016). L’éco-épidémiologie-Vers une épidémiologie de la complexité. médecine/sciences, 32(5), 500–505. https://doi.org/10.1051/ medsci/20163205018. Brandon, R. N. (1990). Adaptation and environment. Princeton University Press. Canali, S. (2020). What is new about the exposome? Exploring scientific change in contemporary epidemiology. International Journal of Environmental Research and Public Health, 17 (8), 2879. https://doi.org/10.3390/ ijerph17082879. Canguilhem, G. (2003). Le vivant et son milieu, In: La Connaissance de la vie. Vrin, Paris. Cheng, P. W., & Novick, L. R. (1991). Causes versus enabling conditions. Cognition, 40(1–2), 83–120. https://doi.org/10.1016/0010-0277(91)900 47-8.

FROM EXPOSOME TO PATHOGENIC NICHE. LOOKING …

201

Egger, G., & Swinburn, B. (1997). An “ecological” approach to the obesity pandemic. BMJ , 315(7106), 477–480. https://doi.org/10.1136/bmj.315. 7106.477. Fox Keller, E. (2010). The mirage of a space between nature and nurture. Duke University Press. Giroux, E. (2008). L’épidémiologie entre population et individu: quelques clarifications à partir de la notion de la pensée populationnelle. Bulletin d’histoire et d’épistémologie des sciences de la vie, 1, 37–52. https://doi.org/10.3917/ bhesv.151.0035. Giroux, É. (2021a). L’exposome: entre globalité et précision. Bulletin d’histoire et d’epistemologie des sciences de la vie, 28(2), 119–148. https://doi.org/10. 3917/bhesv.282.0119. Giroux, É. (2021b). L’exposome: vers une science intégrative des expositions?. Lato Sensu, revue de la Société de philosophie des sciences, 8(3), 9–28. https:// doi.org/10.20416/LSRSPS.V8I3.2. Giroux, É. (2021c). Can populations be healthy? Perspectives from Georges Canguilhem and Geoffrey Rose. History and Philosophy of the Life Sciences, 43(4), 1–23. https://doi.org/10.1007/s40656-021-00463-x. Godfrey-Smith, P. (1998). Complexity and the function of mind in nature. Cambridge University Press. Griffiths, P. E., & Gray, R. D. (1994). Developmental systems and evolutionary explanation. The Journal of philosophy, 91(6), 277–304. https://doi.org/10. 2307/2940982. Guchet, X. (2019). De la médecine personnalisée à l’exposomique, Environnement et santé à l’ère des big data. Multitudes, 75(2), 72–80. https://doi. org/10.3917/mult.075.0072. Holick, M. F. (2020). Sunlight, UV radiation, vitamin D, and skin cancer: How much sunlight do we need? Advances in Experimental Medicine and Biology, 1268, 19–36. https://doi.org/10.1007/978-3-030-46227-7_2. Johnson, C. H., Athersuch, T. J., Collman, G. W., Dhungana, S., Grant, D. F., Jones, D. P., ... & Vasiliou, V. (2017). Yale school of public health symposium on lifetime exposures and human health: The exposome; summary and future reflections. Human Genomics, 11, 32. https://doi.org/10.1186/s40 246-017-0128-0. Juarez, P. D., Matthews-Juarez, P., Hood, D. B., Im, W., Levine, R. S., Kilbourne, B. J., ... & Lichtveld, M. Y. (2014). The public health exposome: A population-based, exposure science approach to health disparities research. International journal of environmental research and public health, 11(12), 12866–12895. https://doi.org/10.3390/ijerph111212866. Kickbusch, I. (1989). Approaches to an ecological base for public health. Health Promotion, 4(4), 265–268. https://doi.org/10.1093/heapro/4.4.265.

202

G. PONTAROTTI AND F. MERLIN

Kirk, S. F., Penney, T. L., & McHugh, T. L. (2010). Characterizing the obesogenic environment: The state of the evidence with directions for future research. Obesity reviews, 11(2), 109–117. https://doi.org/10.1111/j.1467789X.2009.00611.x. Krieger, N. (1994). Epidemiology and the web of causation: has anyone seen the spider?. Social Science & Medicine, 39(7), 887–903. https://doi.org/10. 1016/0277-9536(94)90202-x. Krieger, N. (2001). Theories for social epidemiology in the 21st century: an ecosocial perspective. International Journal of Epidemiology, 30(4), 668–677. https://doi.org/10.1093/ije/30.4.668. Lake, A., & Townshend, T. (2006). Obesogenic environments: Exploring the built and food environments. The Journal of the Royal society for the Promotion of Health, 126(6), 262–267. https://doi.org/10.1177/14664240060704. Laland, K. N., Odling-Smee, F. J., Feldman, M. W. (2001). Niche construction, ecological inheritance, and cycles of contingency in evolution. In: Oyama S., Gray R., Griffiths P. (eds), Cycles of contingency: Developmental systems and evolution. MIT Press, Cambridge. Laland, K. N., Uller, T., Feldman, M. W., Sterelny, K., Müller, G. B., Moczek, A., ... & Odling-Smee, J. (2015). The extended evolutionary synthesis: its structure, assumptions and predictions. Proceedings of the Royal Society B: Biological Sciences, 282(1813), 20151019. https://doi.org/10.1098/rspb.2015.1019. Last, J. M. (1995). A dictionary of epidemiology. Oxford University Press, New York/Oxford/Toronto. Lewontin, R. (1983). The organism as the object and subject of evolution. In: Levins R., Lewontin R. (eds), The dialectical biologist. Harvard University Press, Cambridge. Lewontin, R. C. (2001). The Triple helix. Harvard University Press, Cambridge, MA. Lioy P. J., & Rappaport, S. (2011). Exposure science and the exposome: An opportunity for coherence in the environmental health sciences. Environmental Health Perspectives, 119(11), A-466–7. https://doi.org/10.1289/ ehp.1104387. Mackie, J. L. (1965). Causes and conditions. American Philosophical Quarterly, 2(4), 245–264. https://www.jstor.org/stable/20009173. March, D., & Susser, E. (2006). The eco-in eco-epidemiology. International journal of epidemiology, 35(6), 1379–1383. https://doi.org/10.1093/ije/ dyl249. McLeroy, K. R., Bibeau, D., Steckler, A., & Glanz, K. (1988). An ecological perspective on health promotion programs. Health Education Quarterly, 15, 351–377. Miller, G. W. (2021). Exposome: A new field, a new journal. Exposome, 1(1). https://doi.org/10.1093/exposome/osab001.

FROM EXPOSOME TO PATHOGENIC NICHE. LOOKING …

203

Miller, G. W., & Jones, D. P. (2014). The nature of nurture: Refining the definition of the exposome. Toxicological sciences, 137 (1), 1–2. https://doi.org/ 10.1093/toxsci/kft251. Morris, J. N. (1955). Uses of epidemiology. British Medical Journal, 2(4936), 395–401. https://doi.org/10.1136/bmj.2.4936.395. Niedzwiecki, M. M., Walker, D. I., Vermeulen, R., Chadeau-Hyam, M., Jones, D. P., & Miller, G. W. (2019). The exposome: molecules to populations. Annual review of pharmacology and toxicology, 59, 107–127. https://doi.org/ 10.1146/annurev-pharmtox-010818-021315 . Odling-Smee F. J., Laland K. N., & Feldman M. W. (2003). Niche construction: The neglected process of evolution. Princeton University Press, Princeton. Odling-Smee, J. F. (2010). Niche inheritance. In: Pigliucci M., Müller G. B. (eds), Evolution: The Extended Synthesis (pp. 175–208). Cambridge MA: MIT Press. Parra, R., Cadena, E., & Flores, C. (2019). Maximum UV index records (2010– 2014) in Quito (Ecuador) and its trend inferred from remote sensing data (1979–2018). Atmosphere, 10(12), 787. https://doi.org/10.3390/atmos1 0120787. Patel, C. J., Bhattacharya, J., & Butte, A. J. (2010). An environment-wide association study (EWAS) on type 2 diabetes mellitus. PloS one, 5(5), e10746. https://doi.org/10.1371/journal.pone.0010746. Pearce, T. (2014). The origins and development of the idea of organismenvironment interaction”. In: Barker G. et al. (eds). Entangled life, history, philosophy and theory of the life sciences. vol 4. Springer. https://doi.org/10. 1007/978-94-007-7067-62. Pearl, J. (2009). Causality. New York: Cambridge University Press. Pocheville, A. (2015). The Ecological Niche: History and Recent Controversies In: Heams T., Huneman P, Lecointre G. Silberstein M. (eds), Handbook of evolutionary thinking in the sciences (pp. 547–586). Springer. Pontarotti, G., Dussault, A. C. & Merlin, F. (2022). Conceptualizing the environment in natural sciences: Guest editorial. Biological Theory, 17 , 1–3. https://doi.org/10.1007/s13752-021-00394-7. Rappaport, S. M. (2011). Implications of the exposome for exposure science. Journal of Exposure Science & Environmental Epidemiology, 21(1), 5–9. https://doi.org/10.1038/jes.2010.50. Rappaport, S. M., Barupal, D. K., Wishart, D., Vineis, P., & Scalbert, A. (2014). The blood exposome and its role in discovering causes of disease. Environmental Health Perspectives, 122(8), 769–774. https://doi.org/10.1289/ehp. 1308015. Rappaport, S. M., & Smith, M. T. (2010). Epidemiology. Environment and disease risks. Science (New York, N.Y.), 330(6003), 460–461. https://doi. org/10.1126/science.1192603.

204

G. PONTAROTTI AND F. MERLIN

Raubenheimer, D., Simpson, S. J., & Mayntz, D. (2009). Nutrition, ecology and nutritional ecology: toward an integrated framework. Functional Ecology, 4–16. https://www.jstor.org/stable/40205497. Robinson, O., Tamayo, I., De Castro, M., Valentin, A., Giorgis-Allemand, L., Hjertager Krog, N., ... & Basagaña, X. (2018). The urban exposome during pregnancy and its socioeconomic determinants. Environmental Health Perspectives, 126(7), 077005. https://doi.org/10.1289/EHP2862. Rose, G. (1985). Sick individuals and sick populations. International Journal of Epidemiology, 14(1), 32–38. https://doi.org/10.1093/ije/14.1.32. Sallis, J. F., Owen, N., & Fisher, E. B. (2008). Ecological models of health behavior. In: Glanz K., Rimer B. K., Viswanath K. (eds), Health behavior and health education: Theory, research, and practice (pp. 465–485). Jossey-Bass. Simpson, S. J., Le Couteur, D. G., James, D. E., George, J., Gunton, J. E., Solon-Biet, S. M., & Raubenheimer, D. (2017). The geometric framework for nutrition as a tool in precision medicine. Nutrition and Healthy Aging, 4(3), 217–226. https://doi.org/10.3233/NHA-170027. Siroux, V., Agier, L., & Slama, R. (2016). The exposome concept: A challenge and a potential driver for environmental health research. European Respiratory Review, 25(140), 124–129. https://doi.org/10.1183/16000617.00342016. Stefanovic, N., Irvine, A. D., & Flohr, C. (2021). The role of the environment and exposome in atopic dermatitis. Current Treatment Options in Allergy, 8(3), 222–241. https://doi.org/10.1007/s40521-021-00289-9. Sterelny, K. (2005). Made by each other: Organisms and their environment. Biology and Philosophy, 20(1), 21–36. https://doi.org/10.1007/s10539-0040759-0. Stotz, K. (2017). Why developmental niche construction is not selective niche construction: And why it matters. Interface Focus, 7 (5), 20160157. https:// doi.org/10.1098/rsfs.2016.0157. Stokols, D. (1992). Establishing and maintaining healthy environments: Toward a social ecology of health promotion. American Psychologist, 47 (1), 6–22. https://doi.org/10.1037/0003-066X.47.1.6. Stokols, D., Allen, J., & Bellingham, R. L. (1996). The social ecology of health promotion: Implications for research and practice. American Journal of Health Promotion, 10(4), 247–251. 4 https://doi.org/10.4278/0890-117110.4.247. Susser, M., & Susser, E. (1996). Choosing a future for epidemiology: II. From black box to Chinese boxes and eco-epidemiology. American Journal of Public Health, 86(5), 674–677. https://doi.org/10.2105/AJPH.86.5.674. Swinburn, B., Egger, G., & Raza, F. (1999). Dissecting obesogenic environments: the development and application of a framework for identifying

FROM EXPOSOME TO PATHOGENIC NICHE. LOOKING …

205

and prioritizing environmental interventions for obesity. Preventive Medicine, 29(6), 563–570. https://doi.org/10.1006/pmed.1999.0585. Toscano, W. A., & Oehlke, K. P. (2005). Systems biology: New approaches to old environmental health problems. International Journal of Environmental Research and Public Health, 2(1), 4–9. https://doi.org/10.3390/ijerph200 5010004. Townshend, T., & Lake, A. (2017). Obesogenic environments: Current evidence of the built and food environments. Perspectives in Public Health, 137 (1), 38–44. https://doi.org/10.1177/1757913916679860. Turner, M. C., Vineis, P., Seleiro, E., Dijmarescu, M., Balshaw, D., Bertollini, R., ... & Kogevinas, M. (2018). EXPOsOMICS: Final policy workshop and stakeholder consultation. BMC Public Health, 18(1), 1–11. https://doi.org/ 10.1186/s12889-018-5160-z. Uexk¨ull, J. (1956/1965). Mondes animaux et monde humain. Editions Denoël, Paris. van Tongeren, M., & Cherrie, J. W. (2012). An integrated approach to the exposome. Environmental Health Perspectives, 120(3), a103–a104. https:// doi.org/10.1289/ehp.1104719. Vrijheid, M. (2014). The exposome: A new paradigm to study the impact of environment on health. Thorax, 69(9), 876–878. https://doi.org/10.1136/ thoraxjnl-2013-204949. Vrijheid, M., Slama, R., Robinson, O., Chatzi, L., Coen, M., Van den Hazel, P., ... & Nieuwenhuijsen, M. J. (2014). The human early-life exposome (HELIX): Project rationale and design. Environmental Health Perspectives, 122(6), 535–544. https://doi.org/10.1289/ehp.1307204. Vrijheid, M., Fossati, S., Maitre, L., Márquez, S., Roumeliotaki, T., Agier, L., ... & Chatzi, L. (2020). Early-life environmental exposures and childhood obesity: An exposome-wide approach. Environmental Health Perspectives, 128(6), 067009. https://doi.org/10.1289/EHP5975. Wall, M. M, Larson, N.I., Forsyth, A., Van Riper, D. C., Graham, D. J., Story, M. T., Neumark-Sztainer, D. (2012). Patterns of obesogenic neighborhood features and adolescent weight: a comparison of statistical approaches. American Journal of Preventive Medicine, 42(5), e65–75. https://doi.org/10. 1016/j.amepre.2012.02.009. PMID: 22516505; PMCID: PMC3380614. Walsh, D. (2022). Environment as abstraction. Biological Theory, 17 (1), 68–79. https://doi.org/10.1007/s13752-020-00367-2. West, M. J., & King, A. P. (1987). Settling nature and nurture into an ontogenetic niche. Developmental Psychobiology: The Journal of the International Society for Developmental Psychobiology, 20(5), 549–562. https://doi.org/10. 1002/dev.420200508.PMID: 3678619.

206

G. PONTAROTTI AND F. MERLIN

Wild, C. P. (2005). Complementing the genome with an “exposome”: the outstanding challenge of environmental exposure measurement in molecular epidemiology. Cancer Epidemiology and Prevention Biomarkers, 14(8), 1847–1850. https://doi.org/10.1158/1055-9965.EPI-05-0456 . Wild, C. P. (2012). The exposome: From concept to utility. International Journal of Epidemiology, 41(1), 24–32. https://doi.org/10.1093/ije/ dyr236. Wilding, S., Ziauddeen, N., Smith, D. et al. (2020). Are environmental area characteristics at birth associated with overweight and obesity in school-aged children? Findings from the SLOPE (Studying Lifecourse Obesity PrEdictors) population-based cohort in the south of England. BMC Med 18, 43. https:// doi.org/10.1186/s12916-020-01513-0. Woodward, J. (2003). Making things happen: A theory of causal explanation. Oxford: Oxford University Press. Woodward, J. (2016, Winter). Causation and manipulability. In: Edward N. Zalta (ed), The Stanford Encyclopedia of Philosophy. https://plato.stanford. edu/archives/win2016/entries/causation-mani/.

The Case of Exposome Research: Practical and Disciplinary Issues

Place of Integrative Approaches in the Study of Spatial Dimension of Health Outcomes Yohan Fayet

“Whoever wishes to investigate medicine properly, should proceed thus: in the first place to consider the seasons of the year, and what effects each of them produces for they are not at all alike, but differ much from themselves in regard to their changes. Then the winds, the hot and the cold, especially such as are common to all countries, and then such as are peculiar to each locality. We must also consider the qualities of the waters, for as they differ from one another in taste and weight, so also do they differ much in their qualities. In the same manner, when one comes into a city to which he is a stranger, he ought to consider its situation, how it lies as to the winds and the rising of the sun; for its influence is not the same whether it lies to the north or the south, to the rising or to the setting sun. These things one ought to consider most attentively, and concerning the waters which the inhabitants use, whether they be marshy and soft, or hard, and running

Y. Fayet (B) Human and Social Sciences Department, Centre Léon Bérard, Lyon, France e-mail: [email protected] UMR Inserm U1290 RESHAPE, Lyon, France UMR 5600 Environnement, Ville, Société, Lyon, France

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 É. Giroux et al. (eds.), Integrative Approaches in Environmental Health and Exposome Research, https://doi.org/10.1007/978-3-031-28432-8_8

209

210

Y. FAYET

from elevated and rocky situations, and then if saltish and unfit for cooking; and the ground, whether it be naked and deficient in water, or wooded and well-watered, and whether it lies in a hollow, confined situation, or is elevated and cold; and the mode in which the inhabitants live, and what are their pursuits, whether they are fond of drinking and eating to excess, and given to indolence, or are fond of exercise and labor, and not given to excess in eating and drinking”. On Airs, Waters, and Places, Hippocrates

In this guideline to physicians settling and discovering a new area of practice, Hippocrates (460–370 B.C.) provides a striking insight into his vision of what we now call “the determinants of health” and the potential interactions between environment and health. Insisting on local particularities in terms of physical environment characteristics or social behaviors, he also implicitly recognizes the existence of a spatial variation of health status whose origin would be complex and multifactorial. This concept of the role of the environment on pathology was taken up by Galen, another Greek physician (131–201 B.C.), whose encyclopedia favored the diffusion of Hippocratic thought. If the concept of exposome is nowadays emphasized for its integrative virtues (Lioy and Rappaport 2011), we can see that this holistic vision of pathogenesis was already shared by some illustrious and ancient figures of medicine. The question is not so much whether this holistic and integrative vision of the exposome represents a real novelty in the field of environmental health studies, but rather to determine the extent to which this concept and the techniques associated with it are really contributing to the set-up of a more integrative and holistic knowledge of the environmental determinants of health. The challenge of this chapter will therefore be to situate the historical and scientific contexts in which these “new integrative approaches” are inserted and what their potential contributions or consequences may be. It will be based on a multidisciplinary epistemological analysis (spatial epidemiology, health geography, public health) of past and contemporary research studying the spatial dimension of health and the interactions between health and environment.

PLACE OF INTEGRATIVE APPROACHES IN THE STUDY …

211

1 The Progressive Exploration of Spatial Determinants of Health 1.1

From Early Spatial Analysis to the New Geography of Health

As we can see with Hippocratic thinking (Hippocrates 2021) as well as with the exposome in the scope of the development of precision medicine, some great periods of medical progress have been followed by significant reflection and contributions to the understanding of exogenous factors, and in particular spatial factors, of health. If the Renaissance will be especially marked by the progress in anatomy and in the mechanical understanding of the body, it is at the end of the eighteenth and at the beginning of the nineteenth centuries that the subject of the spatial distribution of the diseases is going to interest again, in particular thanks to the “diffusion of the information on the diseases, the nutrition, the food and the geographical variations in the medical systems” during the Lumières (Earickson 2009). This recognition of the spatial dimension of health status is first achieved thanks to cartography and topography. In 1792, the Prussian Leonhard Ludwig Finke produced a world map of diseases, considered to be the first work of medical geography carried out on a global scale (Light 1944). The revival of Hippocratic thought, thanks to the rediscovery of ancient texts, also inspired certain physicians interested in the relationship between health and place. Through the exercise of medical topography, they carried out a cross description of the living environment and the epidemiological characteristics of the populations. In 1786, Jean-Jacques Menuret de Chambaud wrote an essay on the medical-topographical history of Paris. Initially descriptive, these topographies gradually moved towards a causal approach, thus giving rise to the first works of medical geography. They were particularly useful for understanding and analyzing new epidemics, caused by the first Industrial Revolution and the deterioration of the living conditions of workers in urban areas. In his topography published in 1822, Lachaise related the anemia of certain workers, stricken by tuberculosis, to the insalubrious conditions of their housing. John Snow’s mapping of the cholera epidemic in the Soho district of London in 1855 allowed him to invoke the role of a polluted water fountain and thus to highlight the waterborne origin of cholera. Although strongly criticized at the time, Snow’s conclusions were validated by the discovery of vibrio by Robert Koch in

212

Y. FAYET

1883. The multiplication of medical topographies and the development of public health during the nineteenth century will reinforce the awareness of the implication of living conditions and more globally of the place on health. This even infused the way the city was organized, since the urbanistic upheavals of the end of the nineteenth century in France, even if they were guided by safety imperatives, were also inspired by hygienism and tackled the insalubrity of certain central districts. Nevertheless, these works remained highly monographic and often remained only at the stage of description, based on facts established in a few places. Thus, the monocausal model used in these empirical studies quickly reaches its limits when it comes to considering the overall contribution of place on health status. In France, Maximilien Sorre (1880–1962) was one of the first to go beyond this limit and to propose a theoretical model aimed at interpreting this global effect of place on health. Specialist in biological and human geography and strongly inspired by the conception of geography of Paul Vidal de la Blache (1845–1918), Sorre envisages geography as an “ecology of man, biological and social” (Picheral 2001), through the concepts of environment, milieu or region. In particular, he invented the concept of the “pathogen complex”, as a set of factors conducive to the development of a disease. He thus provided the conceptual framework demonstrating that a pathogen is a necessary but not sufficient cause, and that the disease needs other specific conditions to develop, both in the host and in its environment. Moving away from Vidal de la Blache’s deterministic approach focused on physical conditions, he also remains attentive to the role of human and social dynamics on geographical environments (Sorre 1947). Sorre’s contribution, which established the scientific basis of medical geography, was taken up by Jacques Meyer May (1896–1975) in the United States, notably through the ecological approach to diseases (May 1958; Akhtar 2003; Browne et al. 2018). The adoption of both an ecological and a systemic approach constitutes the great theoretical advance of medical geography founded by Sorre and May. The ecological approach, inspired by both Vidalian thought and American human ecology, focuses on the study of interactions between humans and their environment, in a relationship that is both dialectical and dynamic. Even if it remains well structured around the epidemiology of diseases on the one hand and the organization of care on the other, medical geography introduces tension with the biomedical model, by focusing more particularly on the spatial processes involved in the construction of health. Rejecting all geographical determinism, this

PLACE OF INTEGRATIVE APPROACHES IN THE STUDY …

213

approach integrates the actions of Man over his own environment and considers health as an outcome of the way in which Man integrates and adapts to his environment. These theoretical advances have shown the effect of place, in all its components, on the outcomes of its population. The spatial anchoring of health is not only the result of the configuration of the physical environment, but must also be considered as the result of the human and social dynamics in a place. For a long time, the geographical analysis of health events was based on international comparisons or on monographs, targeting particular places. With the development of public health, epidemiology and medical geography, the spatial analysis of health status has been increasingly studied since the end of the twentieth century, supported by the rapid improvement in the conditions of production, collection and processing of spatial and health data. And since the object of research is no longer the spatial expression of a disease but the influence of place on health, many geographers will call from the 1980s onwards for a broadening of the field of study, through the foundation of a new geography of health. As geographers Moon and Kearns put it, health geography has shifted from a “preoccupation with the medical world to an increased interest in wellbeing and broader social patterns of health and health care” (Kearns and Moon 2002). It has moved away from the biomedical model, interested in all dimensions of well-being, adopting a critical stance towards the discipline but also towards society, towards social inequalities in health, their persistence and their reinforcement. According to Fleuret and Séchet, the objective is no longer “to study health according to places but to study places with regard to health, health care and health policies” (Fleuret and Séchet 2011). This opening was part of the epistemological shift that geography was undergoing at the time, particularly in France, and which refocused the discipline in the field of social sciences. By focusing on lived space, Frémont had, as early as 1976, moved towards a much more phenomenological and behavioral science, thus taking the opposite direction from purely quantitative geography. It is notably from his work that the concept of territory as a portion of space appropriated by a social group, which projects its own system of representation onto it, will develop (Frémont 1976). This new humanistic approach, focused on Man and society, largely explains the progressive emancipation of health geography from traditional medical geography. Already drawn by Antoine Bailly in the early 1980s (Bailly 1981), this movement took shape in

214

Y. FAYET

the 1990s with the launch of the journal Health and Place in 1995 and the explosion in the number of publications on these new themes, focused on well-being and social models. In his dictionary published in 2001, Picheral defines health geography as “the spatial analysis of disparities in the health of populations and the environmental factors (physical, biological, social, economic and cultural) that help explain these inequalities” (Picheral 2001). This definition gives health geography both descriptive and explanatory functions in relation to spatial inequalities in health, by mobilizing all the risk factors potentially involved in a resolutely integrative approach. This integrative approach is not specific to health geographers, but rather reflects a sensitivity specific to geography in general. According to Claval and Pitte, “geographers are interested in the physical, biological and human aspects of the earth. They therefore mobilize concepts that capture in one movement realities that most researchers carefully isolate” (Claval and Pitte 2001). For a long time, the spatial dimension of health remained at the stage of observation or hypothesis, and was therefore gradually investigated with the development of a dedicated discipline, a social science in its own right with its own theoretical baggage. While the emergence of this new discipline has given rise to several notable developments, such as the development of theoretical and critical reflections, the implementation of this integrative approach in the field of spatial analyses was not always as obvious in epidemiology. 1.2

Place as the Proxy of Missing Social Information at the Individual Level

Analytical epidemiology is the study of the distribution and determinants of health and disease in human populations and the causes that determine this distribution. The goal is to prove the relationship between a disease and a risk factor by comparing the frequency of a disease in a group of people exposed to a suspect agent to that in a group of unexposed people. Epidemiology has primarily thought of this relationship at the individual level in order to be able to better control the risk for each individual and to be able to answer the question, “Why did this individual get this disease at this time?”. Thus, ecological studies generally have little consideration in epidemiology because they do not allow the results to be transposed with certainty to the individual level. Indeed, the objective of these studies

PLACE OF INTEGRATIVE APPROACHES IN THE STUDY …

215

is to observe variations in the occurrence of a disease in space (geographical study) or in time (temporal study) and to correlate these variations with environmental factors, in order to develop hypotheses on potential risk factors for the occurrence of pathologies. However, it is difficult to extrapolate inferences made at the (aggregate) population level to the individual level because of what is called the “ecological fallacy”. Given this devaluation of the ecological approach in epidemiology, spatial analysis is generally used as a last resort, in particular to compensate for the lack of information at the individual level. The place was then used as a proxy, a way to measure the characteristics of the individual studied. This approach has been used in particular for social epidemiology studies, in order to observe the influence of social level on health. As medical data still do not often provide information on the social level of patients, spatial data and analysis will be mobilized in order to overcome the great difficulty of collecting social data at the individual level. This work led to the creation of composite indicators to measure social disadvantage. The choice of variables for the indicator obviously depends on the definition of social disadvantage. According to Townsend, social disadvantage is “an observable and demonstrable state of relative disadvantage with respect to the local community or society as a whole to which the individual, family and group belong” (Townsend 1987). The use of these ecological indicators emerged in the United Kingdom in the 1980s, and then spread to most industrialized countries, particularly in the Anglo-Saxon countries (Krieger et al. 2002). The Townsend index was developed in Great Britain in 1987 in the context of the publication of the Black report and it made it possible to show the correlation between deprivation and various health variables (premature mortality, prevalence of chronic diseases, birth weight of newborns) (Black et al. 1980). This famous report on health inequalities in Great Britain insists in particular on the social origin of health inequalities. Socioeconomic index have also been produced in other geographical contexts (Salmond et al. 1998) or with some differences in terms of choice of variables (Jarman 1983; Carstairs and Morris 1989). Initially focused on material deprivation, the method of constructing these indicators will gradually open up to new dimensions of deprivation. In the United Kingdom, the Index of Multiple Deprivation was developed in the 2000s to take into account a larger number of dimensions of deprivation than the strictly socioeconomic indices of Townsend and

216

Y. FAYET

Carstairs. First created in 2004, and then regularly updated, the indicator is based on different areas of disadvantage: income, employment, health, education, access to services, residential environment and crime (Noble et al. 2006). In Canada, the Pampalon Deprivation Index is based on the Townsend model and complements it by integrating the social dimension of disadvantage (Pampalon et al. 2009). The idea is not to combine the two dimensions (material and social) into a single synthetic variable, but to take into account the singularity of each disadvantage. Unlike previous indicators that aim to measure the social vulnerabilities of a territory through a single variable, the final objective of the Pampalon indicator is to allow a more complete mapping of social vulnerabilities, which makes it a highly appreciated tool for health policymakers. Similarly, the DANDEX deprivation indicator developed in Denmark has two components, one measuring social disadvantage (average income, level of education or unemployment rate in the municipalities), the other dealing with material disadvantage with variables on the level of equipment in the territories, housing or car ownership. When compared with mortality data, the Danish indicator makes it possible to observe a “visible gradient” in mortality according to deprivation quintiles (Meijer et al. 2013). The recognition of the role of the life context, alongside individual health determinants, brings us back to the debate in the human and social sciences on the role of collective dynamics on the individual, with the opposition between holistic and individualistic approaches. On the one hand, some argue that the experience of living and sharing the same environment shape individuals and their behavior. This idea of a downward causality from the environment to the individual is based on the work of great authors in the social sciences. On the other hand, the supporters of an individualist approach associate lifestyle and behaviors with a purely personal choice, and therefore independent of the context in which individuals live. As a result, risk is much more individually determined than socially determined from their point of view, so that all variables should be measured at the individual level, because the role of the individual would be much more important in explaining disease (Diez-Roux 1998). This conception of the primacy of individual analysis has long been marked by Robinson’s famous work in sociology pointing out the limits of the ecological approach (Robinson 1950). Macintyre also links this supremacy of methodological individualism to the rise of liberalism in the 1980s, quoting Margaret Thatcher: “There is no such thing as society, there are only individuals”. Macintyre also emphasizes the “important distinction”

PLACE OF INTEGRATIVE APPROACHES IN THE STUDY …

217

between indicators that use aggregate data for individual proxy purposes and those that analyze the effects of the social and physical environment on health (Macintyre et al. 2002). To what extent is the health of a territory due to the characteristics of the individuals who compose it and the context in which the individuals live? In 2002, Shaw recognized that only a joint reading of compositional and contextual effects could provide a complete explanation of health outcomes (Shaw et al. 2002). The development of multilevel models will progressively make it possible to address this complexity and to definitively legitimize the relevance of an integrative and contextual approach to health determinants, by complementing the factors observed at the individual level. 1.3

Multilevel Studies and the Broad Exploration of Contextual Determinants of Health

The interest of multilevel models, including both individual data and spatial variables measuring the characteristics of the life context of individuals, is to be able to distinguish what, in the state of health of an individual, can be linked to his or her own characteristics from what relates to his or her environment (Ellen et al. 2001). This mix within the same statistical model is essential to understand and recognize the importance of contextual effects, which were contested by the biomedical paradigm. Indeed, as Duncan pointed out, until the use of multilevel models, there remained a problem in the interpretation of contextual effects and their importance (Duncan et al. 1998). These studies therefore provided the methodological tools necessary to recognize the influence of the territory as such on health, independently of any other factor or interpretation bias. Of the forty-seven studies included in Riva’s literature review, forty-three showed a significant correlation between the socioeconomic level of the area and one of the different health outcomes, independently of individual characteristics (Riva et al. 2007). Another review of the literature on multilevel analyses focused on mortality, published by Meijer, shows the significant influence of the socioeconomic characteristics of place on health, after control on individual social characteristics (Meijer et al. 2012). Meijer also notes that these effects of place are all the more pronounced when the number of inhabitants of the areas studied is low, which shows for the author the importance of working on small spatial units, at the local level. This effect

218

Y. FAYET

of place, beyond individual characteristics, can be explained on the one hand by “the mutual influence of the inhabitants on the health behaviors of each of them, through the exchange of norms, values and social sanctions” (Meijer et al. 2012). The 2000s were marked by a gradual increase in the number of contextual studies using multilevel models. Whereas ecological studies could give rise to controversy in terms of the interpretation of results, the growing use of multilevel models will greatly contribute to the recognition of the legitimacy and relevance of spatial analysis in health, as a complement to the study of risk factors at the individual level. Alongside the use of social deprivation indices in the framework of multilevel models, the renewed interest in spatial analysis and the study of environmental factors in health will be reflected in the production of geographical indicators measuring new characteristics of the physical environment, beyond the localized measurement of exposure to air, water or soil pollutants. This broadening of the field of investigation of the physical environment has notably favored the recognition of environmental determinants of some health behaviors (diet, alcoholism, smoking), whereas the responsibility for these behaviors is attributed, in a far too exclusive way, to individuals and their non-compliance with prevention messages. Thus, several studies at the turn of the 2010s report that the spatial accessibility of supermarkets, for example, is associated with a higher consumption of fruits and vegetables (Zenk et al. 2009), and globally with a more balanced diet (Larson et al. 2009), confirming the evidence that the proximity of food stores constitutes an important determinant of dietary behaviors and obesity (Holsten 2009; Chaix et al. 2012a). Conversely, the proximity of fast-food outlets or alcohol and tobacco sales outlets is associated with an overconsumption of products sold in these facilities and known for their harmful impact on health. In an urban context, studies have focused on the impact of road quality and urban planning on walking (Roux et al. 2007) and have led to the production of walkability indices, measured by softwares and geographic information systems, particularly in large Anglo-Saxon cities. It should be noted that this broadening of the field of study of the physical environment also allows for a better evaluation of the combined influence of these different determinants. For example, in a 2011 review of the literature, Leal and Chaix already show that some characteristics of the physical environment, such as less pollution, better facilities in terms of shops and services or the presence of facilities that encourage walking,

PLACE OF INTEGRATIVE APPROACHES IN THE STUDY …

219

are all associated with a reduction in risk factors for obesity, hypertension and diabetes (Leal and Chaix 2011). These new studies, which are more integrative in terms of measures of the physical environment, also allow us to reconsider the links between physical and social environments. For example, Cummins’ research on the neighborhood food environment has shown a significant correlation between the presence of McDonald’s restaurants and the level of deprivation in communities in England and Scotland (Cummins et al. 2005). Other studies also show the greater presence of alcohol and tobacco outlets in the most socially deprived areas (Marashi-Pour et al. 2015; Shortt et al. 2015, 2018). The development of contextual and multilevel studies, on a growing number of health outcomes or risk factors, has thus considerably broadened the scope of geographic characteristics studied, to the point that they would be difficult today to list exhaustively. As a consequence of this integrative scientific dynamic, the multiplicity of the studied place effects and the richness of the results progressively contribute to improve the understanding of these health effects of place. Indeed, highlighting the multiplicity and complexity of the effects of place on health has shed a new light to social indicators, until now often used in a restrictive way as a proxy. Finally, this progress supports a probably more exhaustive vision of the environmental factors influencing health outcomes, even if this amount of epidemiological knowledge cannot be sufficient to analyze and explain precisely the mechanism leading to spatial inequalities in health.

2 Integrative Approaches and Methods for the Analysis of Spatial Inequalities in Health The improvement of data collection, management and analysis tools for both health and geographic data, as well as the possibilities offered by Geographic Information Systems (GIS) in the processing and mapping of a large number of data, have greatly facilitated the description and analysis of spatial inequalities in health, through the production of maps at different scales. As a process of significant differentiation of health outcomes according to places, spatial inequalities in health constitute a specific research item that, unlike the work presented above, requires much more (if not almost exclusively) the use of holistic theoretical models and integrative measurement tools.

220

Y. FAYET

2.1

Measurement and Interpretation of Spatial Inequalities in Health

Comparative study of spatial inequalities in life expectancy since the nineteenth century in France has shown that these inequalities are not recent and that they also evolve according to geographical dynamics (Salem et al. 2000). Thus, the departments in the north of France had lower mortality rates than the rest of France at the beginning of the nineteenth century, whereas they have today the lowest life expectancies in metropolitan France. The economic and social crisis initiated by the deindustrialization of these areas, the weight of environmental and occupational exposures, as well as the persistence of some health risk behaviors may explain this relative deterioration (in comparison of the rest of the country) over two centuries. Conversely, we observe a continuous improvement over the last century in the departments of southeastern France. The latest maps of these mortality data confirm the persistence of excess mortality in the North-East of France and in Brittany, but also underline some new dynamics such as the deterioration of the health situation in Languedoc-Roussillon (South of France) (Vigneron and Cartier 2011). Schematically, spatial inequalities in health in metropolitan France reflect different geographic gradients and dynamics on three different scales: first, on a regional scale (Hauts-de-France, Brittany, Grand Est with unfavorable indicators compared to the Occitanie or Provence-AlpesCôte d’Azur regions, for example), between metropolitan areas and rural margins following a center-periphery model, and finally on an urban scale according to the level of social deprivation of neighborhoods. In addition to the description of the spatial distribution of these inequalities, maps using age- and sex-standardized rates can also be used to understand the extent of spatial inequalities in health, by analyzing the difference between the values of the extreme classes on each map. In the case of an atlas of spatial inequalities in health according to the cantons (supra-municipal level) of Metropolitan France, mapping general mortality (all causes of death) according to five color classes, the standardized rates of the class most affected by general mortality are at least twice as large as those of the least affected class (Trugeon et al. 2006). This means that if all French cantons had identical populations, in terms of size and structure by age and sex, we would still observe twice as many deaths per year in some cantons as in others. It is difficult, in the face of such a discrepancy, to implore a random distribution of mortality and to deny

PLACE OF INTEGRATIVE APPROACHES IN THE STUDY …

221

the impact of environmental characteristics in the construction of these health disparities. Above all, the comparative analysis of maps carried out on different pathologies but using the same methodology in the same geographical area also allows us to observe that these differences vary according to the health results used. In the same French atlas (Trugeon et al. 2006), it can be noted that, for cancers of the upper aerodigestive tract or alcohol-related diseases, the rates between the two extreme classes are not multiplied by two, as in the case of general mortality, but by three or even four. Conversely, the differences are smaller for certain diseases such as breast cancer. These variations in the intensity of spatial inequalities mean that the impact of spatial determinants, and therefore the relevance of their analysis, varies according to the diseases. Although they may sometimes seem too descriptive, these maps are further proof of the spatial dimension of health inequalities because they can provide a concrete view of environmental influences on health and help to raise awareness, particularly among public actors, of the importance of these spatial health inequalities and their interconnections with other local issues. The diversity and multiplicity of spatial health inequalities according to the health variables used show the extent to which these inequalities are rooted in specific local differences in living standards, medical infrastructures, environmental exposures and health behaviors. While the spatial analysis of health outcomes has often led to the study of the specific influence of some environmental risk factors on specific health variables, spatial inequalities in health should be considered as the materialization of the unequal geographical distribution of a set of environmental factors that can act positively or negatively on health. 2.2

Spatial Inequalities in Health as the Result of a Holistic Process

Spatial inequalities in health cannot be summarized by a single statistical correlation between one or more environmental characteristics and a health variable, as in the case of the spatial analyses traditionally deployed in health, but are characterized by the multiplicity and complexity of the factors involved in a holistic process of spatial differentiation of health outcomes. This specificity of spatial inequalities in health is not necessarily easy to identify because it does not necessarily correspond to a well-established disciplinary division. Indeed, health geographers can also adopt this analytical approach, measuring the statistical correlation

222

Y. FAYET

between health criteria and territorial characteristics. Nor does the distinction concern the use of different tools, as we have seen with the example of geographical indicators, whether social or environmental, which can be used both to measure an individual’s exposure and to characterize territories. The work of Geoffrey Rose can probably help us to establish this subtle specificity of the analysis of spatial inequalities in health. Indeed, in parallel with the development of spatial analyses in epidemiology, some began to express some criticism of the focus on individual risk factor of health. As early as the mid-1980s, Rose advocated awareness of the specificity of a population-based approach to health, which would not answer the same questions as the individual approach traditionally used. According to Rose, it is not the same thing to look for the causes of cases or the causes of incidence for the same pathology: “‘Why do some individuals have hypertension?’ is quite a different question from ‘why do some populations have hypertension, while it is rare in others?’ These questions require different types of studies, and they have different answers” (Rose 1985). Rose clearly calls for a distinction to be made between etiological research, which focuses on the individual, and research into the causes of incidence, which is measured at the population level. Through various examples from Anglo-Saxon public health, he shows how research focused on the explanation of cases has made it possible to identify “individual susceptibilities” but that this research has failed to identify the underlying causes of disparities in incidence. Rose thus argues for a “Population Strategy” to give priority to finding and controlling the causes of incidence. The interest of Rose’s work is to show the intrinsic importance of population studies in health, to the point of making them an object of research in their own right. This claim for a Population Strategy, complementing the individual approach, generally reflects a renewed interest in environmental factors that may explain health disparities beyond individual characteristics (Macintyre et al. 1993). The diversity of methods and indicators used to study the relationship between health and place should not hide their almost unanimous contribution to the identification of environmental “susceptibilities” at the individual level. The major point of differentiation between the study of spatial inequalities in health and the spatial analyses traditionally deployed in health is therefore mainly the nature of the mechanism studied. Spatial analyses involved in identifying environmental susceptibilities at the individual level make it possible to grasp the biological

PLACE OF INTEGRATIVE APPROACHES IN THE STUDY …

223

mechanism by which a social/environmental exposure leads to the appearance of a pathology. In the case of spatial inequalities in health, the aim is to measure, through health data, a social mechanism of differentiation between different populations in terms of social class or geographical context for example. This holistic principle is thus applied to the study of social inequalities in health, comparing different health indicators according to socio-professional categories, and which share the same concern to decipher the social process responsible for health inequalities, by seeking the causes of this differentiated social construction. The analysis of spatial inequalities in health is therefore fundamentally part of a deeply holistic and integrative approach, aiming to consider all the environmental characteristics involved in this process of spatial differentiation of health outcomes. 2.3

Integrative Tools Supported by the Analysis of Spatial Inequalities in Health

The extensive revelation of spatial determinants of health suggests the geographic contexts’ potential to produce health inequalities. However, estimating this impact of the geographic context on health inequalities remains difficult, partly because epidemiological studies aiming to identify associations between spatial characteristics and health outcomes most often investigate one spatial factor at a time, according to the objectives of the study. As a result, geographic contexts may be variously measured, in terms of characteristics and spatial scale. In addition, spatial indices may be combined differently across studies, due to methodological choices and study objectives. Some international collaborations have aimed to develop methodologies for standard indices using spatial data, as in the case of the European Deprivation Index (Pornet et al. 2012; Launoy et al. 2018). However, for many geographic characteristics, data availability and scale are too variable between and within countries to develop the set of standard indicators needed to measure and compare spatial health inequalities in a consistent way. Consequently, studies analyzing spatial inequalities in health often used specific indices, limiting the comparability of the results (Abel et al. 2016). Furthermore, the separated analysis of risk factors in epidemiological studies mostly impedes a comprehensive review of all the vulnerabilities related to the place of residence. Some epidemiological studies

224

Y. FAYET

using a social deprivation index took this limit into account, investigating a potential difference in their analysis between rural and urban deprived areas (Bertin et al. 2014). Considering this challenge, developing geographical classification (or typology) can help to summarize all the geographical determinants on health in a meaningful way and to develop a common geographical frame of reference for the study of spatial inequalities in health. The use of geographic classifications to compare health outcomes is relatively recent (Gershoff et al. 2009), but some significant examples can be mentioned worldwide. Using data on the characteristics of the physical environment (built environment and housing) and the population (social level, ethnic origin and communities, demography), Weden creates a spatial typology for the study of health inequalities throughout the United States (Weden et al. 2011). In the end, the typology distinguishes six “archetypes” of territories and observes their evolution between 1990 and 2000. The authors insist on the stability of the numbers between 1990 and 2000, which proves the viability of this typology over time and the capacity of the model to measure the temporal evolution of geographic health inequalities. Still in the United States, Arcaya uses 55 variables for its typology of the state of Massachusetts, divided into six different domains: health behaviors, housing and land use, transportation, services, social composition and demographic composition. The authors see this typology as an aid to programming, communication and evaluation of local health policies (Arcaya et al. 2014). In Brazil, Santos presents a five-class typology, based on demographic, social and housing conditions data in 794 microneighborhoods (>5000 inhabitants) in Rio de Janeiro (Santos et al. 2010). The classes are distinguished according to socioeconomic level and urban/rural character, and will be used for future work in health (accidents, violence, communicable diseases and mortality). In France, this comprehensive classification approach to address the geographical context’s contribution into health inequalities has been also used with the production of the “Geographical Classification for Health studies” (Fayet et al. 2020). This classification was computed, through k-means clustering, from ten spatial variables measuring physical environment, social deprivation and healthcare accessibility at the municipality level. The classification distinguishes 5 types of municipalities (Wealthy Metropolitan Areas, Precarious Population Districts, Residential Outskirts, Agricultural and Industrial Plains, Rural Margins) which enables to highlight significant spatial inequalities in standardized

PLACE OF INTEGRATIVE APPROACHES IN THE STUDY …

225

mortality between the 5 classes. Indeed, significant lower mortality rates compared to the mainland France population were found in the Wealthy Metropolitan Areas (SMR = 0·868,95%CI: 0·863–0·873) and in the Residential Outskirts (SMR = 0·971, 95%CI: 0·964–0·978), while significant excess mortality were found for Precarious Population Districts (SMR = 1·037,95%CI: 1·035–1·039), Agricultural and Industrial Plains (SMR = 1·066, 95%CI: 1·063–1·070) and Rural Margins (SMR = 1·042,95%CI: 1·037–1·047). At the level of the Paris metropolitan area, Van Hulst also produces a spatial classification for the analysis of spatial inequalities in cardiovascular disease (Van Hulst et al. 2012). This typology distinguishes six profiles of territories, mainly according to their degree of urbanization and their social composition (Van Hulst 2012). The description of the typology shows, for many characteristics, strong contrasts between the different profiles. These geographic contrasts are expected to result in significant differences in cardiovascular risk exposures, between populations in these different territories. Data from the RECORD (Residential Environment and Coronary Heart Disease) cohort, which measures systolic (maximum pressure at the time of heart contraction) and diastolic (minimum pressure at the time of “relaxation” of the heart) blood pressure in more than 7000 French people living in the Paris metropolitan area, have verified this hypothesis of a differentiated exposure to cardiovascular risk, depending on the type of territory and its characteristics. The results show a significantly higher systolic blood pressure in people living in deprived urban areas, after adjustment for individual risk factors. There was also a consistent decrease in diastolic blood pressure as one moved away from urban centers. Finally, at equivalent urban density, the social disadvantage of the neighborhood also influences the increase in diastolic blood pressure and thus the cardiovascular risk. This study highlights the relevance of an integrative spatial approach to observe inequalities in cardiovascular risk. At the end of the paper, Van Hulst highlights that “the typology makes it possible to examine the combined exposure to multiple environmental characteristics that are highly correlated and whose effects could not be separated through multivariable regression analysis. By regrouping similar neighborhoods based on a multidimensional profile, it is possible to examine the impact of a constellation of neighborhood environment features that may jointly rather than individually influence health and health behaviors” (Van Hulst et al. 2012).

226

Y. FAYET

Published in 2012, Van Hulst’s paper presents a precursory example of a study combining an integrative spatial approach and the use of biological data to measure the incorporation of socio-spatial inequalities in health. At a time when these same virtues are being emphasized for exposome supporters, it is appropriate to evaluate the impact of these new exposomic studies in the field of research studying the spatial dimension of health events.

3 Exploring Health and Place Through the Exposome: Opportunities, Knowledge and Challenges Exposomic studies aim to use biomarkers to trace the life-course effects of environment on health, following an integrative approach of exposures. It represents therefore a significant opportunity for better integration of environmental measures into health studies with high level of precision, thanks to molecular data. The exposome is claimed to improve knowledge about environmental factors impacting health outcomes at the individual level, taking cumulative effects of different exposures over the life course into account. Raising the benefit of the concept of exposome for health geography, Prior also points that “the holistic nature of the exposome is particularly beneficial to the integration of biosocial ideas into geographic health enquiry” (Prior et al. 2019). Indeed, “biosocial theorisations enable both body and environment to be repositioned as active components in fluid health and place relationships, acting in interchange and accumulation over time” (Prior et al. 2019). However, the high precision of the biological data collected in these cohorts very often comes up against the scarcity of spatial data measuring exhaustively and/or precisely the exposure to different environmental factors, whether physical or social. Even if conditions for collecting and sharing spatial data measuring these environmental characteristics are gradually improving, these imbalances in terms of precision between biological data and environmental data, and even between the environmental data themselves, call into question the ability of the exposomic approach to actually implement its integrative ambition.

PLACE OF INTEGRATIVE APPROACHES IN THE STUDY …

3.1

227

What About the Integration of Spatial Data in Exposomic Studies?

The first large studies concretely implementing the exposomic approach were born in the early 2010s, within the framework of major international collaborations bringing together different cohorts on the same subject. This first generation of large exposomic studies is characterized by its marked focus on the analysis of biomarkers associated with certain physical environmental exposures, in particular to air pollutants. For example, the HELIX (Human Early Life Exposome) study was initiated in 2013 to describe the multiple environmental exposures during pregnancy and childhood, in order to identify associations between these early exposures, “molecular signatures” and diseases in children (Maitre et al. 2018). Bringing together different European cohorts on neonatal health (9 regions and 6 countries for a total of 31 472 women), HELIX integrates the measurement of 17 different exposures (climate, air pollution, built environment, chemical agents, etc.). Nevertheless, the study only integrates the geographical areas (region, metropolitan area, etc.) for which data on air pollution and built environment are available (Maitre et al. 2018), which shows the importance of data availability in the design of the studies and the factors considered. Another example is the EXPOsOMICS study, funded by the European Union and involving 13 European and American research centers “to develop a novel approach to the assessment of exposure to high priority environmental pollutants” and which is clearly focused “on air and water contaminants during critical periods of life” (Vineis et al. 2017). Note that within this first generation of exposomic studies, only the LIFEPATH study does not incorporate measures of physical environment exposures and instead focuses on the biological effects of the social environment on health aging, from 8 cohorts in France, Italy, Portugal, Ireland, the United Kingdom, Finland, Switzerland and Australia (Vineis et al. 2017). According to Prior, “the lack of the social is damaging to exposomic studies” because “environmental exposures and their biological correlates cannot be separated from the broader social, economic, political and cultural relations in which they are embedded” (Prior et al. 2019). This unbalanced integration of spatial variables and factors in the first exposomic studies could be explained by the greater availability and better precision of some data on physical environments (mainly air pollution),

228

Y. FAYET

that are routinely produced by national and/or local institutions in European countries. One can also think that air pollution, which can be measured by a “simple” concentration rate of particles, was a good model for the implementation of an exposomic approach correlating measurements of internal and external exposomes. While the measure of some other risk factors of physical environment (e.g. accessibility of green/blue spaces or facilities, walkability) was not often routinely produced and required the construction of more complex indicators, air pollution data were ready to use and to be integrated into early exposomic studies. Launched in 2020, the European Human Exposome Network (EHEN) brings together 9 research projects on exposome, receiving over e100 million from Horizon 2020, the EU’s framework program for research and innovation. Presenting itself as the world’s largest network of projects studying the impact of environmental exposure on human health, this EHEN seems to support a second generation of exposomic studies, more integrative especially in terms of spatial data. Indeed, even if air pollution data are still used, most of the nine projects more or less aim to include some other spatial data measuring physical (e.g. food/alcohol outlets; urban land uses, population density, walkability, green/blue spaces, climate, odor, noise) or social environments (e.g. lifestyle and behaviors, income, social capital or networks). While most of them clearly display their commitment to a holistic and integrative approach of exposures, this ambition is more visible and advanced for some projects in particular, such as Expanse (Vlaanderen et al. 2021), Athlete (Vrijheid et al. 2021) and Equal-life (Kamp et al. 2022). However, the EHEN projects do not seem to be able to grasp the geographical diversity of the health and place interactions. Indeed, several projects now mention or even clearly assume a specialization of their studies on urban exposome. Moreover, most of the spatial variables integrated measuring physical environments (walkability, accessibility of food/alcohol outlets, accessibility of green/blue spaces) are mostly designed for urban context. Knowing that studies may choose not to include, in their analyses, patients for whom spatial data would not be available (Maitre et al. 2018), one can therefore ask the question of the representation of non-urban health issues.

PLACE OF INTEGRATIVE APPROACHES IN THE STUDY …

3.2

229

Novelties and Challenges of the Exposomic Approach for the Study of Spatial Dimension of Health Outcomes

The concept of exposome is not only presented as an interesting opportunity but often as a real innovation, a turning point bringing a major change in the study of interactions between health and environment (Rappaport 2018; Vineis 2018; Vineis et al. 2020). However, this chapter reported previous experiences and progress in the study of spatial dimension of health outcomes that are not without interactions with the holistic and integrative approach claimed by the exposome concept. This narrative on the alleged novelty of the exposome has already been criticized (Canali 2020). We have already shown that the integrative approach to health determinants was already widely advocated in other disciplines, such as health geography (see 1.1), and that it was also necessary for the understanding and analysis of some issues such as socio-spatial inequalities in health (see 2.2). We have also seen that multilevel analyses also aim to bring together all environmental exposures and individual susceptibilities within a single causal model in order to estimate the influence of each of these factors. Some of these studies even used biomarkers to quantify the effects of different exposures on health through a process of biological incorporation. This is the case, for example, of the RECORD study which, since the end of the 2000s, has combined biological measurements, socio-demographic information at the individual level and data measuring the physical and social environment in order to study the impact of the residential environment on coronary heart disease (Chaix et al. 2012b; Van Hulst et al. 2012). Presenting final results of the EXPOsOMICS study, Turner and his/her co-authors recognize that “on a basic level, exposome research can be seen as replicating the approaches of classic risk assessment with higher resolution and greater accuracy” (Turner et al. 2018). The allegedly “holistic” or integrative dimension of exposome would be related for some to the several exposures taking into account at the same time, making it possible to examine “the effects of multiple classes of agents as part of a more holistic approach to risk assessment” (Turner et al. 2018). Actually, the exposome seems rather to extend and intensify a holistic and integrative scientific dynamic that already existed, for various reasons, in the field of spatial analyses in health. Even if former studies combined biological, individual and environmental data in a life-course perspective, what is striking about these exposomic studies is the scale at which they

230

Y. FAYET

are implemented. These studies often take shape around large international consortia bringing together numerous scientific disciplines, pooling their data, their tools and their study populations. The exposome thus seems to be the concept capable of federating a scientific community, originally plural, around a single approach and of supporting a massification of studies on environmental health. These advances in data integration, however, require exposomic studies to deal more with variability in data accuracy and measurement tools. While they benefit from the greater precision of biological measurements, studies have more difficulties in routinely collecting precise data on the socio-demographic and residential trajectories of individuals, but also on their current and past exposures. While exposomic studies incorporate more exposures of different types, there is a risk of combining data from various measures whose robustness, quality and completeness may be very uneven (see Giroux in this volume). The limited availability and accuracy of precise spatial data, compared to individual and biological data, therefore represent a major obstacle and challenge for the continuation and balance of exposomic studies. Indeed, these problems lead teams to make choices that are sometimes contradictory with the integrative ambition of the exposome, such as excluding patients for whom spatial data are not available (Maitre et al. 2018) or the non-inclusion of cohorts not collecting some spatial data. Thus, the included cohorts in the Lifepath project “represent only a small proportion of the total cohorts available in Europe” because they had to combine “good measures of socioeconomic status, risk factors for noncommunicable diseases and biomarkers already measured” (Vineis et al. 2017). While the EHEN projects benefit from significant political and financial support from the EU Commission, massive investment in tools for routine measurement and collection of environmental exposures and characteristics could also have been very useful considering the limited availability and accuracy of spatial data. The improvement of environmental measurement tools is all the more important as the temptation to identify and use biomarkers to measure exposures directly in the body becomes stronger and stronger, in a movement of biologization of sociospatial inequalities whose risks are still far from being controlled (Lynch 2017; Serviant-Fine et al. 2023). The integrative virtues of the exposome thus fortuitously highlight the significant imbalances between the tools and data grouped within

PLACE OF INTEGRATIVE APPROACHES IN THE STUDY …

231

exposomic studies. This reality is well taken into account by scientists, as evidenced by certain methodological choices, and these current constraints could foster an awareness of the need for investment in environmental data measurement and production tools. While waiting for a potential improvement in data homogeneity that would allow for an even more integrative implementation of the exposome concept, we can observe that exposome studies are mostly implemented in urban or even metropolitan environments and based on direct data on exposure to pollution. Some environmental factors or areas, initially considered in the cohorts, may also be excluded from the analyses in the end, due to the variable availability of precise spatial data. It can be seen here that the integrative ambition of exposomic studies leads them in practice to a spatial and factorial reductionism which may question their claimed capacity to respond to major public health issues. In France, for example, a recent study shows that some metropolitan areas stand out for their high undermortality, even though these areas are the most exposed nationwide to air pollution (Fayet et al. 2020). Exposomic studies may therefore increase socio-spatial inequalities in health since it will produce useful knowledge to assess the impact of urban exposome and to support policies reducing exposures in the densest urban areas at the expense of others areas. Moreover, socio-spatial inequalities in health are mainly linked to social environment and health-related behaviors (tobacco, alcohol, diet), which are currently significantly underrepresented in these studies.

4

Conclusion

Considered at different times in the spatial analysis of health outcomes and now necessary for the analysis and understanding of socio-spatial inequalities in health, the holistic and integrative approach to interactions between health and place is undoubtedly experiencing a new dynamic under the effect of exposomic research. The intensification of this integrative dynamic allowed by the exposome concept is also characterized by the massification of studies, reaching unprecedented scales in terms of infrastructures, study populations, data and disciplines mobilized thanks to the rise of big data. However, the exhaustive and precise measure of environmental factors potentially contributing to health outcomes and inequalities is still limited by technical and financial constraints which question the representativeness of the studies and their ability to address all public health issues, usually reported by studies in epidemiology and

232

Y. FAYET

health geography. This should lead us to qualify not the scientific interest of the exposome but its claim to provide objective knowledge to support policies addressing public health issues, such as socio-spatial inequalities in health.

References Abel, Gary A, Matthew E Barclay, and Rupert A Payne. 2016. Adjusted indices of multiple deprivation to enable comparisons within and between constituent countries of the UK including an illustration using mortality rates. BMJ Open 6. https://doi.org/10.1136/bmjopen-2016-012750. Akhtar, Rais. 2003. Medical geography: Has J.M. May borrowed M. Sorre’s 1933 concept of pathogenic complexes? Cybergeo: European Journal of Geography. CNRS-UMR Géographie-cités 8504. https://doi.org/10.4000/cyb ergeo.3976. Arcaya, Mariana, Timothy Reardon, Joshua Vogel, Bonnie K Andrews, Wenjun Li, and Thomas Land. 2014. Tailoring community-based wellness initiatives with latent class analysis–Massachusetts Community Transformation Grant projects. Preventing Chronic Disease 11: E21. https://doi.org/10.5888/ pcd11.130215. Bailly, Antoine S. 1981. La géographie du bien-être. Espace et liberté, ISSN 0222–3376 5. Paris, France: Presses universitaires de France. Bertin, Mélanie, Cécile Chevrier, Fabienne Pelé, Tania Serrano-Chavez, Sylvaine Cordier, and Jean-François Viel. 2014. Can a deprivation index be used legitimately over both urban and rural areas? International Journal of Health Geographics 13: 22. https://doi.org/10.1186/1476-072X-13-22. Black, Douglas, Jerry Morris, Cyril Smith, and Peter Townsend. 1980. Inequalities in Health: The Black Report. Department of Health and Social Security (DHSS). Browne, Dillon T., Mark Wade, Heather Prime, and Jennifer M. Jenkins. 2018. School Readiness amongst Urban Canadian families: Risk profiles and family mediation. Journal Of Educational Psychology 110: 133–146. https://doi. org/10.1037/edu0000202. Canali, Stefano. 2020. What is new about the exposome? Exploring scientific change in contemporary epidemiology. International Journal of Environmental Research and Public Health 17: 2879. https://doi.org/10.3390/ije rph17082879. Carstairs, Vera, and Russel Morris. 1989. Deprivation: Explaining differences in mortality between Scotland and England and Wales. BMJ: British Medical Journal 299: 886–889. Chaix, Basile, Kathy Bean, Mark Daniel, Shannon N Zenk, Yan Kestens, Hélène Charreire, Cinira Leal, et al. 2012a. Associations of supermarket characteristics

PLACE OF INTEGRATIVE APPROACHES IN THE STUDY …

233

with weight status and body fat: A multilevel analysis of individuals within supermarkets (RECORD study). PloS one 7: e32908. https://doi.org/10. 1371/journal.pone.0032908. Chaix, Basile, Yan Kestens, Kathy Bean, Cinira Leal, Noëlla Karusisi, Karima Meghiref, Julie Burban, et al. 2012b. Cohort profile: Residential and nonresidential environments, individual activity spaces and cardiovascular risk factors and diseases–the RECORD Cohort Study. International Journal of Epidemiology 41: 1283–1292. https://doi.org/10.1093/ije/dyr107. Claval, Paul, and Jean-Robert Pitte. 2001. Épistémologie de la géographie. Paris, France: Nathan. Cummins, Steven C. J., Laura McKay, and Sally MacIntyre. 2005. McDonald’s restaurants and neighborhood deprivation in Scotland and England. American Journal of Preventive Medicine 29: 308–310. https://doi.org/10.1016/j.ame pre.2005.06.011. Diez-Roux, Anna V. 1998. Bringing context back into epidemiology: Variables and fallacies in multilevel analysis. American Journal of Public Health 88: 216–222. Diez-Roux, Ana V., Kelly R. Evenson, Aileen P. McGinn, Daniel G. Brown, Latetia Moore, Shannon Brines, and David R. Jacobs. 2007. Availability of recreational resources and physical activity in adults. American Journal of Public Health 97: 493–499. https://doi.org/10.2105/AJPH.2006.087734. Duncan, C, K Jones, and G Moon. 1998. Context, composition and heterogeneity: Using multilevel models in health research. Social science & medicine (1982) 46: 97–117. Earickson, R. 2009. Medical Geography. In International Encyclopedia of Human Geography, ed. Editors-in-Chief: Rob Kitchin and Nigel Thrift, 9–20. Oxford: Elsevier. Ellen, Ingrid Gould, Tod Mijanovich, and Keri-Nicole Dillman. 2001. Neighborhood Effects on Health: Exploring the Links and Assessing the Evidence. Journal of Urban Affairs 23: 391–408. https://doi.org/10.1111/07352166.00096. Fayet, Y., Delphine Praud, Béatrice Fervers, Isabelle Ray-Coquard, Jean-Yves Blay, Françoise. Ducimetiere, Guy Fagherazzi, and Elodie Faure. 2020. Beyond the map: Evidencing the spatial dimension of health inequalities. International Journal of Health Geographics 19. https://doi.org/10.1186/ s12942-020-00242-0. Fleuret, Sébastien, and Raymonde Séchet. 2011. Spatialité des enjeux de pouvoir et des inégalités: pour une géographie sociale de la santé. In Penser et faire la géographie sociale : Contributions pour une épistémologie de la géographie sociale, 333–351. Rennes: Presses Universitaires de Rennes.

234

Y. FAYET

Frémont, Armand. 1976. La région, espace vécu. 1 vols. SUP, Le géographe [Texte imprimé]/section dir. par Pierre George. - Paris: Presses universitaires de France, 1968–1983 19. Paris, France: Presses universitaires de France. Gershoff, Elizabeth T., Sara Pedersen, and J. Lawrence Aber. 2009. Creating neighborhood typologies of GIS-based data in the absence of neighborhoodbased sampling: A factor and cluster analytic strategy. Journal of Prevention & Intervention in the Community 37: 35–47. https://doi.org/10.1080/108 52350802498458. Hippocrates. 2021. Airs, Waters, Places. Good Press. Holsten, Joanna E. 2009. Obesity and the community food environment: A systematic review. Public Health Nutrition 12: 397–405. https://doi.org/ 10.1017/S1368980008002267. Jarman, Brian. 1983. Identification of underprivileged areas. British Medical Journal (Clinical Research Ed.) 286: 1705–1709. https://doi.org/10.1136/ bmj.286.6379.1705. Kamp, Irene van, Kerstin Persson Waye, Katja Kanninen, John Gulliver, Alessandro Bozzon, Achilleas Psyllidis, Hendriek Boshuizen, et al. 2022. Early environmental quality and life-course mental health effects: The EqualLife project. Environmental Epidemiology 6: e183. https://doi.org/10.1097/ EE9.0000000000000183. Kearns, Robin, and Graham Moon. 2002. From medical to health geography: Novelty, place and theory after a decade of change. Progress in Human Geography 26: 605–625. https://doi.org/10.1191/0309132502ph389oa. Krieger, Nancy, Jarvis T. Chen, Pamela D. Waterman, Mah-Jabeen Soobader, S. V. Subramanian, and Rosa Carson. 2002. Geocoding and monitoring of US socioeconomic inequalities in mortality and cancer incidence: Does the choice of area-based measure and geographic level matter?: The Public Health Disparities Geocoding Project. American Journal of Epidemiology 156: 471– 482. Larson, Nicole I., Mary T. Story, and Melissa C. Nelson. 2009. Neighborhood environments: Disparities in access to healthy foods in the U.S. American Journal of Preventive Medicine 36: 74–81. https://doi.org/10.1016/j.ame pre.2008.09.025. Launoy, Guy, Launay, Ludivine, Dejardin, Olivier, Bryère, Joséphine, and Guillaume, Elodie. 2018. European Deprivation Index: Designed to tackle socioeconomic inequalities in cancer in Europe. European Journal of Public Health 28. https://doi.org/10.1093/eurpub/cky213.625. Leal, Cinira, and Chaix, Basile. 2011. The influence of geographic life environments on cardiometabolic risk factors: A systematic review, a methodological assessment and a research agenda. Obesity Reviews: An Official Journal of the International Association for the Study of Obesity 12: 217–230. https://doi. org/10.1111/j.1467-789X.2010.00726.x.

PLACE OF INTEGRATIVE APPROACHES IN THE STUDY …

235

Light, Richard Upjohn. 1944. The Progress of Medical Geography. Geographical Review 34: 636. https://doi.org/10.2307/210033. Lioy, Paul J., and Stephen M. Rappaport. 2011. Exposure science and the exposome: An opportunity for coherence in the environmental health sciences. Environmental Health Perspectives 119: A466–467. https://doi.org/10. 1289/ehp.1104387. Lynch, Julia. 2017. Reframing inequality? The health inequalities turn as a dangerous frame shift. Journal of Public Health 39: 653–660. Oxford, England. https://doi.org/10.1093/pubmed/fdw140. Macintyre, Sally, Anne Ellaway, and Steven Cummins. 2002. Place effects on health: How can we conceptualise, operationalise and measure them? Social Science & Medicine 55: 125–139. https://doi.org/10.1016/S02779536(01)00214-3. Macintyre, Sally, Sheila Maciver, and Anne Sooman. 1993. Area, class and health: Should we be focusing on places or people? Journal of Social Policy 22: 213– 234. https://doi.org/10.1017/S0047279400019310. Maitre, Lea, Jeroen de Bont, Maribel Casas, Oliver Robinson, Gunn Marit Aasvang, Lydiane Agier, Sandra Andrusaityte, et al. 2018. Human Early Life Exposome (HELIX) study: A European population-based exposome cohort. BMJ open 8: e021311. https://doi.org/10.1136/bmjopen-2017-021311. Marashi-Pour, Sadaf, Michelle Cretikos, Claudine Lyons, Nick Rose, Bin Jalaludin, and Joanne Smith. 2015. The association between the density of retail tobacco outlets, individual smoking status, neighbourhood socioeconomic status and school locations in New South Wales, Australia. Spatial and Spatio-Temporal Epidemiology 12: 1–7. https://doi.org/10.1016/j.sste.2014. 09.001. May, Jacques Meyer. 1958. The Ecology of Human Disease. MD Publications. Meijer, Mathias, Gerda Engholm, Ulrike Grittner, and Kim Bloomfield. 2013. A socioeconomic deprivation index for small areas in Denmark. Scandinavian Journal of Public Health 41: 560–569. https://doi.org/10.1177/140 3494813483937. Meijer, Mathias, Jeannette Röhl, Kim Bloomfield, and Ulrike Grittner. 2012. Do neighborhoods affect individual mortality? A systematic review and metaanalysis of multilevel studies. Social Science & Medicine 74: 1204–1212. https://doi.org/10.1016/j.socscimed.2011.11.034. Noble, Michael, Gemma Wright, George Smith, and Chris Dibben. 2006. Measuring multiple deprivation at the small-area level. Environment and Planning A 38: 169–185. https://doi.org/10.1068/a37168. Pampalon, R., D. Hamel, P. Gamache, and G. Raymond. 2009. A deprivation index for health planning in Canada. Chronic Diseases in Canada 29: 178– 191.

236

Y. FAYET

Picheral, Henri. 2001. Dictionnaire raisonné de géographie de la santé. 1 vols. Montpellier, France: Université Montpellier III-GEOS. Pornet, Carole, Cyrille Delpierre, Olivier Dejardin, Pascale Grosclaude, Ludivine Launay, Lydia Guittet, Thierry Lang, and Guy Launoy. 2012. Construction of an adaptable European transnational ecological deprivation index: The French version. Journal of Epidemiology and Community Health 66: 982–989. https://doi.org/10.1136/jech-2011-200311. Prior, Lucy, David Manley, and Clive E. Sabel. 2019. Biosocial health geography: New ‘exposomic’ geographies of health and place. Progress in Human Geography 43: 531–552. https://doi.org/10.1177/0309132518772644. Rappaport, Stephen M. 2018. Redefining environmental exposure for disease etiology. NPJ Systems Biology and Applications 4: 30. https://doi.org/10. 1038/s41540-018-0065-0. Riva, Mylene, Lise Gauvin, and Tracie A Barnett. 2007. Toward the next generation of research into small area effects on health: A synthesis of multilevel investigations published since July 1998. Journal of Epidemiology and Community Health 61: 853–861. https://doi.org/10.1136/jech.2006. 050740. Robinson, W. S. 1950. Ecological correlations and the behavior of individuals. American Sociological Review 15: 351. https://doi.org/10.2307/2087176. Rose, G. 1985. Sick individuals and sick populations. International Journal of Epidemiology 14: 32–38. Salem, Gérard, Stéphane Rican, Eric Jougla, Cyrille Suss, and Marianne BerthodWurmser. 2000. Atlas de la santé en France. Volume 1, Les causes de décès. Collection MIRE, ISSN 1621–1014. Montrouge : J. Libbey Eurotext. Salmond, Clare, Peter Crampton, and Frances Sutton. 1998. NZDep91: A New Zealand index of deprivation. Australian and New Zealand Journal of Public Health 22: 835–837. https://doi.org/10.1111/j.1467-842x.1998. tb01505.x. Santos, Simone M., Dora Chor, and Guilherme L. Werneck. 2010. Demarcation of local neighborhoods to study relations between contextual factors and health. International Journal of Health Geographics 9: 34. https://doi.org/ 10.1186/1476-072X-9-34. Serviant-Fine, Thibaut, Mathieu Arminjon, Yohan Fayet, and Élodie Giroux. 2023. Allostatic load: historical origins, promises and costs of a recent biosocial approach. BioSocieties. https://doi.org/10.1057/s41292-023-003 03-0. Shaw, Mary, Daniel Dorling, Richard Mitchell, and Peter Haggett. 2002. Health, place, and society. Harlow, England, Royaume-Uni. Shortt, Niamh K., Esther Rind, Jamie Pearce, Richard Mitchell, and Sarah Curtis. 2018. Alcohol risk environments, vulnerability and social inequalities

PLACE OF INTEGRATIVE APPROACHES IN THE STUDY …

237

in alcohol consumption. Annals of the American Association of Geographers 108: 1210–1227. https://doi.org/10.1080/24694452.2018.1431105. Shortt, Niamh K., Catherine Tisch, Jamie Pearce, Richard Mitchell, Elizabeth A. Richardson, Sarah Hill, and Jeff Collin. 2015. A cross-sectional analysis of the relationship between tobacco and alcohol outlet density and neighbourhood deprivation. BMC Public Health 15: 1014. https://doi.org/10.1186/ s12889-015-2321-1. Sorre, Maximilien. 1947. Les fondements de la géographie humaine. Paris, France: Colin. Townsend, Peter. 1987. Deprivation. Journal of Social Policy 16: 125–146. https://doi.org/10.1017/S0047279400020341. Trugeon, Alain, Danièle Fontaine, Bernadette Lémery, Xavier Bertrand, and Fédération nationale des observatoires régionaux de santé. 2006. Inégalités socio-sanitaires en France: de la région au canton. Issy-les-Moulineaux, France: Masson, DL 2006. Turner, Michelle C., Paolo Vineis, Eduardo Seleiro, Michaela Dijmarescu, David Balshaw, Roberto Bertollini, Marc Chadeau-Hyam, et al. 2018. EXPOsOMICS: Final policy workshop and stakeholder consultation. BMC Public Health 18: 260. https://doi.org/10.1186/s12889-018-5160-z. Van Hulst, Andraea, Frédérique Thomas, Tracie A. Barnett, Yan Kestens, Lise Gauvin, Bruno Pannier, and Basile Chaix. 2012. A typology of neighborhoods and blood pressure in the RECORD Cohort Study. Journal of Hypertension 30: 1336–1346. https://doi.org/10.1097/HJH.0b013e3283544863. Vigneron, Emmanuel, and Nicolas (1975) Cartier. 2011. Les inégalités de santé dans les territoires français : état des lieux et voies de progrès. Issy-les-Moulineaux: Elsevier Masson. Vigneron, Emmanuel, and Cartier, Nicolas. 2011. Les inégalités de santé dans les territoires français : état des lieux et voies de progrès. Issy-les-Moulineaux: Elsevier Masson Vineis, Paolo. 2018. From John Snow to omics: The long journey of environmental epidemiology. European Journal of Epidemiology 33: 355–363. https://doi.org/10.1007/s10654-018-0398-4. Vineis, Paolo, Marc Chadeau-Hyam, Hans Gmuender, John Gulliver, Zdenko Herceg, Jos Kleinjans, Manolis Kogevinas, Soterios Kyrtopoulos, M Nieuwenhuijsen, and David H. Phillips. 2017. The exposome in practice: design of the EXPOsOMICS project. International Journal of Hygiene and Environmental Health 220: 142–151. Vineis, Paolo, Oliver Robinson, Marc Chadeau-Hyam, Abbas Dehghan, Ian Mudway, and Sonia Dagnino. 2020. What is new in the exposome? Environment International 143: 105887. https://doi.org/10.1016/j.envint.2020. 105887.

238

Y. FAYET

Vlaanderen, Jelle, Kees de Hoogh, Gerard Hoek, Annette Peters, Nicole ProbstHensch, Augustin Scalbert, Erik Melén, et al. 2021. Developing the building blocks to elucidate the impact of the urban exposome on cardiometabolicpulmonary disease: The EU EXPANSE project. Environmental Epidemiology (Philadelphia, Pa.) 5: e162. https://doi.org/10.1097/EE9.000000000000 0162. Vrijheid, Martine, Xavier Basagaña, Juan R. Gonzalez, Vincent W. V. Jaddoe, Genon Jensen, Hector C. Keun, Rosemary R. C. McEachan, et al. 2021. Advancing tools for human early lifecourse exposome research and translation (ATHLETE): Project overview. Environmental Epidemiology (Philadelphia, Pa.) 5: e166. https://doi.org/10.1097/EE9.0000000000000166. Weden, Margaret M., Chloe E. Bird, José J. Escarce, and Nicole Lurie. 2011. Neighborhood archetypes for population health research: Is there no place like home? Health & Place 17. Health Geographies of Voluntarism: 289–299. https://doi.org/10.1016/j.healthplace.2010.11.002. Zenk, Shannon N, Laurie L Lachance, Amy J Schulz, Graciela Mentz, Srimathi Kannan, and William Ridella. 2009. Neighborhood retail food environment and fruit and vegetable intake in a multiethnic urban population. American Journal of Health Promotion: AJHP 23: 255–264. https://doi.org/10.4278/ ajhp.071204127.

The Exposome and the Social Sciences: The Case of Systemic Diseases Catherine Cavalin

1 Introduction: The Post-Genomic Historical Moment 1.1

About the Complex Aetiology of Disease

As a preface to the first part of his popular science book on the history and issues of genetics, Siddhartha Mukherjee quotes a dialogue from Oscar Wilde’s play The Importance of Being Earnest. “Jack: Yes, but you said yourself that a severe chill was not hereditary. Algernon: It usen’t to be, I know—but I daresay it is now. Science is always making wonderful improvements in things.” (Mukherjee 2016, p. 16). To caricature this amusing exchange, one could say that: the more we learn from science, the greater the likelihood that we will base our ills on genetic mechanisms. We are almost twenty years after the announcement of the complete sequencing of the human genome (Human Genome

C. Cavalin (B) National Centre for Scientific Research (CNRS), Université Paris-Dauphine (PSL), Paris, France e-mail: [email protected]

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 É. Giroux et al. (eds.), Integrative Approaches in Environmental Health and Exposome Research, https://doi.org/10.1007/978-3-031-28432-8_9

239

240

C. CAVALIN

Project), and many genome-wide association studies have developed in the wake of this research project of unprecedented scope on a global scale. The goal of genome-wide association studies has been to determine statistically significant correlations, on the scale of the largest and most representative population samples possible, between genetic variations and the various phenotypes of numerous diseases. From the Wildian playlet, let’s place ourselves in an alternate historical perspective. What would our present be if those scientific discoveries on the human genome had, in the last two decades, explained all diseases? Or again: what would our present be if those works of the last twenty years had brought even the promise of such exhaustive explanations? Whatever the precise scenario that might have resulted from such a miracle, one can at least hypothesize that the social sciences of health would be in a subordinate position to the life sciences. In the case of a bijective relationship between a given phenotype of a specific disease and a given nucleotide polymorphism, there would indeed be no room left to explain—outside of genes—the difference between two individuals’ health statuses, whether a sick and a non-sick person, or even two sick people differently affected by the severity of the same disease. It is precisely because we are now in a very different situation that the exposome concept was born. And it is precisely because we are now in this very different situation that there are many reasons to engage in an interdisciplinary discussion, beyond the gap between social and biomedical sciences. In a way, the social sciences as well as the exposome approach share the objective of “complementing” the genome, as suggested by the first word of Christopher P. Wild’s seminal article. As genomics do not give us all the explanations for diseases (chronic and non-infectious diseases, in particular), the post-genomic era is both made of “euphoria and disappointment” (Senier et al. 2017, 108): there is still a lot to do, but there are hopes and promises for all disciplinary knowledge on health. 1.2

The Social Sciences at the Door of the Exposome?

Social epidemiologist Nancy Krieger (1994) noted that images (and metaphors) play a special role in scientific knowledge production. They help—and sometimes also hinder—to conceive fresh ideas. Since its issuing in 2012, the diagrammatic representation of three different domains of the exposome by Christopher P. Wild (Wild 2012, p. 25)— internal, specific external and general external—opens at least three

THE EXPOSOME AND THE SOCIAL SCIENCES: THE CASE …

241

research paths, the holistic and the integrative dimensions of the exposome being supposedly, as defined by Wild, the outcome of the tight collaboration between those three domains. It is still impossible to assess whether this didactic drawing will persistently keep an influential role in the framing or shaping of epidemiological and biomedical thinking. However, from this little diagram, as well as from many publications claiming to adhere to the exposome approach, we can already perceive various sources of tension and ambiguity both within the exposomic approach (as represented by the three domains of inquiry), and between the exposome and disciplines working in social research fields. In particular, the exposome may be torn between holism and reductionism (Giroux et al. 2021); but it may also be torn between a comprehensive approach of exposure through the whole life course and the development of a precise measurement of all those exposures and the biological mechanisms they are related with (Giroux 2021a); and finally cut across by an ambiguous project of integration (integration of data or exposure variables? exposure and/or effect on health? methods? explanations? disciplines?) (Giroux 2021b, p. 10). All those possible ambiguities or tensions between the various dimensions encompassed by the exposomic approach, which are addressed in other chapters of this book (see Giroux this volume, Pontarotti & Merlin this volume), may seem particularly sensitive to social scientists. Indeed, in all the literature resulting from huge research projects conducted under the umbrella of the exposome, and despite Wild’s definition of a “general external” sphere of environmental determinants of our health (Wild 2012, p. 28), there is little mention of the possible or proven participation of the social sciences in the production of knowledge. And in the major institutional moments gathering the heavyweight stakeholders (researchers or funders) of the exposome, the life sciences also seem to enjoy a great deal of autonomy, particularly regarding the social sciences. As an edifying example, the launching of the European Human Exposome Network (EHEN) (The EU Framework Programme for Research and Innovation et al. 2020) under the patronage of the European Commission thus brought together (around Christopher P. Wild) not only an Areopagus of professionals working in biology, toxicology or (environmental, molecular) epidemiology, but also representatives of European Community administrations responsible for implementing health policies. In this context, would the place and role played by the social sciences be inexistent in the exposome era? If the “exposomers”—as we could call

242

C. CAVALIN

the researchers in biomedical sciences involved in those big “exposomelabelled” research projects—show no interest in the methods and results that the social sciences can bring, would the social sciences abandon dialogue with the exposome? This chapter endeavours to provide a negative answer to such a question. We will support the idea of possible and close relationships between social scientists and exposomers’ research questions, (a) by expanding on the historical reasons for the social sciences and exposomic research not to ignore each other; (b) by reminding that ongoing research on both sides already expresses some forms of collaborations, and that those outstretched hands continue a long history of studying the social determinants of health; (c) by explaining how, in particular, certain research questions related to systemic diseases may drive the social sciences to propose research methods and fieldworks compatible with the exposome; and at last, (d) by developing some examples about interdisciplinary research on those diseases, in order to explain that the collaboration proposed by the social sciences is based on effective empirical tools of inquiry.

2

2.1

How Much Pluralistic and Cooperative Are Current Collaborations Between the Exposome and the Social Sciences? Supplementing and Questioning the Exposome Framing Through the Lens of Social Inequalities in Health

This chapter began by pointing out the potential gap between the exposome and the social sciences. However, on both sides, research initiatives exist (though rare), that are likely or willing to bridge this gap. Even though the exposome is often displayed as an imposing “paradigm” (for instance when Stephen Rappaport defines the exposome as the “operational paradigm” of “exposure science” (Rappaport 2011, p. 5)) by the exposomers themselves, this paradigmatical status is not totally overwhelming and monolithic. Indeed, some initiatives demonstrate the vigour of a certain epistemological and methodological pluralism around the exposome. It is necessary to do justice to those inter- or transdisciplinary efforts, and to recall here briefly some specificities of their contributions, and how fruitful they can be.

THE EXPOSOME AND THE SOCIAL SCIENCES: THE CASE …

243

To get a sense of the pluralistic nature of the discussion of socioenvironmental determinants of health around the exposome, we will first draw on two texts (Juarez et al. 2014; Senier et al. 2017) that can be considered as emblematic because: (i) they take the paradigmatic nature of the exposome so seriously that they themselves make ambitious “exposomic” proposals (through the public health exposome and the socio-exposome, respectively); (ii) beyond their similarities, their differences help to think more generally what plurality can consist of on the common basis of “exposome” thinking, and how the social sciences can contribute to it. The authors of those two articles who formulate their research avenues in terms of “exposome” share a common concern for social inequalities of health (health disparities and health inequalities, respectively). Strictly speaking, all those authors are not “social scientists”, in a narrow disciplinary meaning. Yet, they all develop an approach that is—at least— compatible with social sciences, by their questioning health issues through the lens of social inequalities. Since 2014, Paul D. Juarez has directed the newly formed Research Center on Health Disparities, Equity, and the Exposome (Memphis, TN). Despite the name of this research centre, we must notice that the public health exposome by (Juarez et al. 2014) is conceptualized almost autonomously from the exposomers’ exposome. References to Rappaport (2011) and Wild (2012) come only in the eleventh and eighteenth positions in Juarez et al. (2014)’s bibliography. Significantly, the opening words of the article define the problem to be solved by referring to the issue of socio-racial/ethnic health inequalities, in order to address immediately their possible socio-environmental determinants, and above all the difficulty to capture, assess and measure them. The solution offered to face this difficulty lies in the operationalization of tracking environmental exposures through the public health exposome as a big device including numerous sources of data (particularly socio-spatial data) on “natural, built, social and policy environments”. One of the peculiarities of this device lies in the multiplication of the sources of data, their nature and the numerous levels at which they are captured, having in mind that health disparities must be uncovered through “social-ecological contexts.”1 The public health exposome not only calls various disciplines 1 In the era of “Big Data” (and giving a very interesting definition of “Big Data” (Juarez et al. 2014, p. 12,874)), Juarez et al.’s position reminds us very closely the one defended before this sociotechnical era by (Susser et Susser 1996).

244

C. CAVALIN

(within and without the biomedical sciences) to action. It is also repeatedly presented by its authors as willing to put data into the hands of end users or the hands of community members (Juarez et al. 2014, p. 12,878, p. 12,880). Sharing the data, its production and use with the communities affected by the health problems examined by the research is part of the research itself. On this point, Juarez et al.’s “public health exposome” is partially in accordance with Senier et al.’s “socio-exposome”. Yet, consistently with popular epidemiology and environmental justice (Brown 1997; Brown et al. 2012), the socio-exposome takes one more critical step forward, by stressing even more the importance of lay knowledge and Wild’s oversight (Senier et al. 2017, p. 113) about it. Patients or community members have to be involved in research: they must know what they have been exposed to; they must also participate as full stakeholders in the research, also acting as investigators searching for what they have been exposed to, and suggesting ways to make it “perceptible” (as perceptibility is defined by (Murphy 2006, p. 10)). This position is not only a political criticism of the exposome such as it is conceived by the “exposomers”. It goes hand in hand with STS scholarship of what “exposure” is. Exposure is not granted, nor universally defined. Could the public health exposome and the socio-exposome be considered as forming parts of a same paradigm, that would be competing with the exposome paradigm? Whatever the answer, they are different ways of displaying pluralistic approaches of exposure assessment and health inequalities. Whereas Juarez et al. insist more on the technical stakes of building a data collection device on ecological exposure and population health status, Senier et al. highlight the necessity to pay attention to social and political conditions that keep allowing toxic exposure. Beyond their differences, they challenge the way in which the exposome defines exposures, proposes to evaluate or measure them, tends to neglect external exposures in favour of an internal chemical exposome and also tends to overlook the historical and sociopolitical contexts that make deleterious exposures possible or even legal. 2.2

Biosocial Approaches Based on the Notion of Embodiment: The LIFEPATH Project

The LIFEPATH project directed by Paolo Vineis claims to include social characteristics of individuals in an analysis that seeks to reveal social determinants of health. This epistemological option is generally presented by

THE EXPOSOME AND THE SOCIAL SCIENCES: THE CASE …

245

LIFEPATH’s promoters as one of the hallmarks of this research project. Encompassing data drawn from up to 17 European cohorts (Vineis et al. 2017), LIFEPATH aims to highlight causal mechanisms by which social characteristics and socio-economic contexts produce health effects through the mediation of biological properties. This project certainly places a strong emphasis on the social characterization of individual life paths, and as such at least partly leans towards a holistic view of health, in which the use of social sciences is legitimated. A recent article co-authored with Robert Barouki on the possible contributions of epidemiology and toxicology “under the general umbrella of the ‘exposome’” even presents social factors as ultimate (upstream) determinants of health (“no longer […] a confounding factor but […] an overarching determinant”) (Vineis et Barouki 2022, p. 7). Nevertheless, as Élodie Giroux has shown (Giroux 2021b), the mechanistic model of embodiment implemented by LIFEPATH might not make social or population characteristics interact with biology. The relationship would be causal and one way, from (socially characterized) exposure via—and towards—biological parameters. To understand better the relationships between biological and social in LIFEPATH, interviewing Paolo Vineis and his research teammates on their use of the Bourdieusian notion of “capital” would be highly interesting. With this in mind, we can find quite a nuanced expression of the bio/social interrelationships and interactions in another text co-authored with Michelle Kelly-Irving. Paolo Vineis and Michelle KellyIrving suggest the existence of a somewhat transitive mechanism from social determinants (social inequalities) to unequal health statuses, via the transmission belt of lifestyles and health behaviours: “Disadvantaged socio-economic position (SEP) in early life may shape lifestyle and healthrelated behaviours, which then affect health in adulthood” (Vineis et Kelly-Irving 2019, p. 980). If we rely on this latter formulation, the social would thus be incorporated into health (and in the form of health statuses), through stages that are built upon more or less healthy “lifestyles”. Based on those three approaches, we can observe some epistemological and methodological pluralism, with some variations in the degrees and forms of collaboration between the exposome and the social sciences. However, again here, the social sciences seem to be in a dependent and relatively secondary position.

246

C. CAVALIN

3 Systemic Diseases as a Fruitful Fieldwork to Bridge the Gap Between the Exposome and the Social Sciences 3.1

Starting from Empirical Fieldworks

Obviously, most of the discourses on possible collaborations or crossfertilizations between the exposome and the social sciences come from researchers who are involved in situated (i.e. empirical) fieldworks: “citizen science” to understand the prevalence of some chronic illnesses, “environmental justice” or “environmental racism” in the case of Juarez et al. (2014) or Senier et al. (2017), “sociomarkers” (Ghiara et Russo 2019) as embodiment of biosignatures via social exposures for Vineis and his collaborators. Yet, the part dedicated to general or theoretical (i.e. not empirical) considerations on the bridges, gaps, possibilities to outstretch hands, etc. between exposomers’ and social scientists’ research is huge in those researchers’ reflections. Probably because the exposome being a young and remarkably successful scientific proposal, it keeps fuelling debate on its broadest stakes, and probably also because, as a “technoscientific promise” (Joly 2010), it generates scientific debate in broad and comprehensive terms. I would like to get back to empirical grounds, to discuss epistemological and methodological stakes through practical examples. I can even remind here that, as a sociologist, my personal interest in the exposome has originally been grounded on the necessity to find empirical tools of inquiry. While working with physicians in various disciplines (mainly pulmonology, internal medicine, rheumatology) and with patients affected by systemic diseases, while preparing surveys designed to be conducted in the general population, I had to operationalize ways of uncovering possible “environmental” health determinants. What is the exposome? What can I understand through it? Is it compatible with my own research lines on those fieldworks about health? I questioned myself this way because for me as well as for the exposomers, the postgenomic context mentioned above provides the starting point. We all work on what remains to be explained or understood about some chronic diseases, we all question the part played by “other” health determinants than those studied by genomics, and the possible interactions between the two. Nevertheless, I am not a physician nor a biomedical scientist implementing “omic” approaches. From that, which empirical tools would I

THE EXPOSOME AND THE SOCIAL SCIENCES: THE CASE …

247

need to build, in order to enter what Christopher P. Wild classified as the “general external” domain of what is not gene-determined in health statuses? 3.2

The Complexity of Systemic Diseases Brings Together Issues of Primary Importance to Exposomers and Social Scientists

Summarizing some of the peculiarities of systemic diseases is instrumental to enter the comprehension of the specific form of dialogue with the exposome that my “shoe-leather”2 research endeavours to develop from the social sciences. Systemic diseases are the conditions by which the patients I am working with are affected. Systemic diseases are considered to be “of unknown aetiology”, according to biomedical and epidemiologic current knowledge. Which means that they typically embody the characteristics of the post-genomic research questions: if some genes may explain part of their incidence or some of their phenotypes, they are conceived in medical and epidemiological reasoning as “complex” diseases. As such, they potentially unite “environmental” multiple determinants and individual genetic “susceptibility”, making discussion about their “causes” multilayered and multidisciplinary (Parascandola 2011). In biomedical sciences (including epidemiology), systemic diseases happen to be also described as “plurisystemic” insofar as they involve (or can involve) several organs or even several “systems”, in the sense that medicine speaks, for example, of the vascular, respiratory, blood system, etc. They are most often inflammatory and immune-mediated (dysimmune or autoimmune diseases). Many of them are registered as rare diseases, according to the criterion of one case for a population of 2,000 people, but some are not, the most prevalent of them being rheumatoid arthritis (RA). In total, those diseases which are in particular—in addition to RA—systemic lupus erythematosus (SLE) and systemic sclerosis (SSc), sarcoidosis, dermatomyositis, vasculitis, etc., affect a total of about one and a half million people in France. In the United States, the prevalence of autoimmune diseases alone is estimated to be 3%, and they are one of

2 With a wink to the so-called shoe-leather epidemiology, which refers to a fieldwork intended to uncover health social determinants for populations and individuals. See for instance (Koo et Thacker 2010).

248

C. CAVALIN

the main causes of death in women of median age (Cooper et Stroehla 2003). Not only the systemic diseases’ causes (or cofactors or triggers) appear unknown and require responses in our post-genomic era. Almost more intriguing than their onset and incidence, is indeed the great variability of their phenotypes. Their severity is highly variable, and some of them can either present themselves under the invisible form of asymptomatic conditions or lethal diseases. Sarcoidosis can illustrate this broad spectrum of phenotypes, as it may be diagnosed by chance in a routine radiological check-up—as was the case for young conscripts when military service still existed in France and an x-ray of the lungs was carried out for the purpose of health surveillance against tuberculosis. Yet, sarcoidosis may also be symptomatic (often with respiratory symptoms), and so severe that it can cause the death of people affected, particularly when organs such as heart or kidneys are involved. Between the exposomic approach and an inquiry in the social sciences about the patients’ “environment”, we thus understand that questions on causality but also on the variability of those diseases provide obvious common avenues of research.

4 Interdisciplinary Research on Systemic Diseases: What is at Stake? 4.1

Facing the Intriguing Diversity of Phenotypes: Crossing Tools of Inquiry, Comparing Fieldworks

The portrait of systemic diseases is thus intriguing, and their diagnosis is often very difficult and takes much time to elaborate. In this context, the diversity of phenotypes is particularly characterized by a very unbalanced sex ratio in patients: the vast majority (80% for SLE, more than 70% for RA or SSc) of patients are female, and when men are affected, it is most often severely. Similarly, for many of those diseases, the incidence and severity of cases is more marked according to race, with more frequent and more severe pathologies among people of African or subSaharan origin. Despite these striking features, there is no actual social epidemiology of these diseases. Some studies are just beginning to show how individual characteristics (sex, race), which have long been treated as risk factors in themselves, i.e. as “biological” properties, actually refer

THE EXPOSOME AND THE SOCIAL SCIENCES: THE CASE …

249

to a social characterization. In the case of SSc, for example, serological markers and differential mortality indicate greater severity of SSc in African-American patients than in those of other backgrounds. But once the effect of household income level is controlled, the “effect” of race vanishes (Moore et al. 2019). And beyond characteristics of sex or race, the form and severity of these diseases are, for all patients, very difficult to predict, even once a diagnosis is made by a team of expert clinicians. This point is crucial because these diseases cannot be cured. Treatments (e.g. corticosteroids, immunosuppressants) have dreadful side effects, and can only slow down the course of the disease. Therefore, one of the physicians’ most pressing challenges is to formulate the most proactive and accurate prognosis possible in order to choose the nature and timing of a therapy in an individually tailored way, for the long-term benefit of the patient. This therapeutical challenge gives convincing motivations for the social sciences to collaborate with physicians’ clinical practice and biomedical sciences researching on systemic diseases. At stake in producing a social epidemiology of systemic diseases is the unveiling of unseen characteristics of the patients likely to help understand the seeming inconsistence of diverse and unpredictable pathological phenotypes. The expected clinical benefit (i.e. a better tailored treatment for a better quality of life and potentially a longer life) is the responsibility of physicians. For their part, the social sciences can hypothesize that those diseases are social facts, and consequently conduct a fieldwork inquiry able to answer the following question: behind the many uncertainties of systemic diseases, where can their social regularities be found? How could patients and their diseases be newly qualified so that a social narrative could give an account of their situation, and help to treat them? Part of the results I have been producing in the last years along with colleagues in the social sciences and biomedical sciences have endeavoured to answer such questions. One of our main instruments has been the building and implementation of a questionnaire both applied to interrogating patients (with possible combination with biographical interviews) and conducting a statistical survey in the general population. This questionnaire has produced unprecedented data exploring exposure to crystalline silica (and secondarily some other inorganic particles) in occupational and non-occupational settings in the whole life course. By

250

C. CAVALIN

comparing patients and their matched3 controls sampled in the general population, this questionnaire has uncovered the great statistical significance of occupational exposure to crystalline silica particles for patients having SSc and RA. This result cannot be interpreted as “precise” in the sense that it would provide a mechanistic explanation of these autoimmune diseases. Yet, this statement about precision deserves several comments. First, the quality of the data on which the highlighting of the occupational dimension of the exposures is based, associated with a deep rereading of the causes in the history of these diseases, encourages the firm identification of types of occupational environments at risk. In this sense, a precision of results can be claimed, although with a different meaning than that claimed by exposomers. Being able to compare patient numbers that are quite large on the scale of these diseases (especially the rarer SSc) with data sampled in the rolling census by the French National Statistical Institute (INSEE) (Cavalin et al. 2021; Cavalin et al. 2022), a control population, whose high quality is not always achieved by epidemiological surveys, helps to consolidate past results. Those results also confirm the possible differential risk for men to be more severely affected by occupational overexposures to inorganic particles, although they are less affected by the incidence of these diseases than women. Furthermore, the implications of correlations between these possible environmental determinants and the diseases of interest should not be underestimated. For a physician specialized in SSc or RA, to see a male patient is a relatively rare occurrence. The possibility of thinking about reconstructing the history of exposure to inorganic particles in his professional past opens up the possibility of better adapting the drug treatment to his disease in the long term and, depending on the social protection system, of giving him access to compensation for occupational disease, for example. The aim is to reconcile the objectives that Geoffrey Rose displayed (Rose 1985) as rarely achievable in a single movement: to unveil and act on “the causes of cases” and “the causes of incidence”. The research here combines methods of interrogation in the social sciences and a clinical perspective, in an effort to improve patient care. Moreover, this approach is reproducible. As the previous study has shed a new light on the gendered phenotypes of SSc and RA, an

3 Matching controls with patients on sex, age range and tobacco consumption criteria.

THE EXPOSOME AND THE SOCIAL SCIENCES: THE CASE …

251

analogous questionnaire (the same as the previous one, with a few adaptations to the specificities of an inquiry with children and their families) has produced novel results on sarcoidosis in children and teenagers. Sarcoidosis, a systemic (dysimmune) disease which is even more rare in children than in adults, is generally very severe when it happens in young people. Among the 36 patients who were interviewed, 24 were of sub-Saharan or Caribbean origin. Our inquiry was based upon the questionnaire, but also upon biographical and family interviews, and associated with a similar survey with random controls as well as young controls suffering from sickle-cell disease. As it could be the case for lead poisoning a few decades ago (Fassin 2003), our results (Nathan et al. 2021) contribute to “deracialize” sarcoidosis or “debiologize” race in this disease. Among other things, indeed, our results do not point at “race” as an ending point risk factor. On the contrary, by comparing children with sarcoidosis and with sickle-cell disease whose families have similar geographical origins, they highlight the statistical significance of overexposure to inorganic particles in occupational setting in the young patients’ families, as well as the role that insalubrious housing conditions could play in the onset of the children’s disease. 4.2

Addressing Causality by Putting Systemic Diseases in Historical Perspective

The way causality is discussed in biomedical sciences and epidemiology in relation to systemic diseases leads to a dialogue between the exposome paradigm and the social sciences by other paths. The example of sarcoidosis, RA and SSc here again provides interesting insights in this respect. The first sentence or paragraph of articles on sarcoidosis almost ritualistically presents this dysimmune granulomatosis as “a systemic disease of unknown cause” (Valeyre et al. 2014), even when research precisely identifies some of the possible causes (or cofactors or triggers) of this disease, particularly via exposure in occupational (Oliver et Zarnke 2021) and “environmental” settings (Newman et Newman 2012). These mysteries about the causality of sarcoidosis, both described as “of unknown aetiology” and the causes of which are documented by tangible evidence, point significantly to three structuring features in long-term research on immune-mediated disease causation. First, it can be hypothesized that the persistent representation in medical knowledge of “causeless” diseases

252

C. CAVALIN

necessarily exerts framing effects (Rosenberg et Golden 1992) on the ontology of these diseases, both for physicians and for patients. Second, the absence of “known” causes goes hand in hand in the medical literature with the description of these diseases as having “complex” causes. In this way, the discourse on these diseases joins the very terms in which exposomic research defines its privileged targets of investigation, i.e. unexplained disease phenotypes, whose causes are accepted as “complex”, in the sense of both “genetic and environmental”. Finally, the mystery of causality in systemic immune-mediated diseases is the legacy of a (still active) history in which certain causes, though long identified as such, tend to disappear periodically in medical knowledge and training. This relative amnesia consists, for example, in not setting in the stone of medical knowledge the possible occupational origin of some of these diseases, which has been identified at certain times. While some of this “forgetfulness” can be explained by the economic interests at stake (as when employers may have an interest in minimizing the occupational causes of certain illnesses in order not to compensate them (Cavalin et al. 2021a, b)), the instability of knowledge also most certainly has more complex causes (as shown by the complexity of the mechanisms by which ignorance can be produced in any form of information or knowledge production (Proctor et Schiebinger 2008)). Taken together, these phenomena make the search for causality on these diseases difficult. In those matters, working in the social sciences—and precisely here in history and sociology of scientific knowledge—can contribute to gather and reconstruct scattered pieces of past knowledge, in order to reflect afresh on current knowledge. That was the underlying epistemological significance of the SILICOSIS project,4 in which the questionnaire mentioned above was initially developed. Rereading the history of silicosis (Rosental et al. 2015) which is a pneumoconiosis i.e. a disease caused by the accumulation of crystalline silica particles in the respiratory system helps open research avenues on analogous possible occupational origins of systemic diseases that do not have always respiratory symptoms. Immunology can take up research in which immune mechanisms relating to those diseases, cancers and pneumoconiosis are discussed; history and sociology can help clinicians to become aware of the historical character 4 ERC Advanced Grant project, directed by Paul-André Rosental, Centre for European Studies and Comparative Politics, Sciences Po, Paris 2012–2017. Grant number ERC2011-ADG_20110406, project ID 295,817.

THE EXPOSOME AND THE SOCIAL SCIENCES: THE CASE …

253

of nosology (Lescoat et al. 2019), and thus help them rethink classifications in an attempt to better manage patients. On the whole, by unearthing and rebuilding the steps of past knowledge, the social sciences can contribute to make interdisciplinary current (and hopefully) future knowledge more firmly grounded, while in the long run it is always likely to remain less cumulative than one could ideally imagine about scientific knowledge. This approach to interdisciplinarity through a labour in history and sociology of scientific knowledge is a manner to actively seek the building of scientific long-lasting “thought collectives” (Fleck 1979) about environmental health. 4.3

Addressing the Very Consistency of Systemic Diseases’ Nosology

Finally, a third line of research will be summarized here to discuss the methods and “precision” claimed by research based on the exposome paradigm. A clinical team specializing in SSc, one of whose members (Alain Lescoat) also co-produced the results mentioned above, has set up a research programme questioning the very nosology of SSc. In the currently accepted nosology of this disease, the “limited” form (which affects about two-thirds of the patients) is usually distinguished from the “diffuse” form, on the basis of criteria of localization and extension of skin involvement. This partition has been retained for what it allows to understand the clinical prognosis, and as such has played a structuring role for the study design in clinical research on SSc. The fact that the limited form is often less severe, however, does not prevent it from being accompanied by a severe prognosis and a significant deterioration in quality of life. And recent literature has focused mainly on the characterization and management of the health status of patients with the diffuse form. Consequently, clinical research is not sufficiently adapted to the most frequent forms of the disease to seek the best therapies and effectively measure the progression of the patients’ health status (Lescoat et al. 2021, 2020). This justifies giving a voice to patients with a limited form of SSc, to have them freely describe their symptoms, the areas of their health on which these have an impact, and their experience of quality of life (Lescoat et al. 2022). The objective is to produce a new composite indicator of characterization of this form of SSc, likely to make the criteria for judging therapeutic research more relevant, for a better benefit of these patients whose case had been neglected in favour of the diffuse form of SSc. All in

254

C. CAVALIN

all, the approach is therefore that of reconstruction, after a stage of deconstruction of the existing nosological categories. This approach makes it possible to discuss both the “precision” and “completeness” claimed by the exposome by proposing alternative paths. Looking for completeness and precision about diseases the nosology of which is not properly designed for therapeutic purposes deserves indeed to be discussed. This clinical research on SSc shows thus the interest of taking an acute look— including that of the patients—at the nosological entities themselves, in a more ecological and social approach of health.

5 Conclusion: Leveraging the Case of Systemic Diseases “of Unknown Aetiology” In the current historical context of a post-genomic moment, the case of systemic diseases seems particularly relevant to open deeper collaborations with the exposome, with proactive avenues for the social sciences. Until now, indeed, the social sciences have rarely managed to legitimate their research proposals, even if the “general external” domain of the exposome may theoretically welcome them. Improving the therapeutic management of patients is specially a challenge that the joint action of mechanistic (exposomic) research and social science surveys can address, united around common research questions on the intriguing uncertainties and diversity of those diseases’ phenotypes. Those common goals do not imply that everything in inter- or transdisciplinary research would be conceived the same way for each discipline involved. The social sciences may produce data of a special relevance and accuracy, offering other forms of “precision” than the exposome science intends to produce. Moreover, behind the unexplained heterogeneity of health statuses, while the exposome has a legitimate role to provide mechanistic insights, the social sciences rather look for social facts, understandable in terms of social inequalities in health. Those differences in interpretation or objectives again can be fruitful, as they can help rethink the very nosology of unexplained diseases: redefine the very outline of diseases to better explain them, and thus make the most of the patients’ experience. On the whole, this chapter’s conclusions agree with the historical approach developed by Michel Dubois, Catherine Guaspare and Séverine Louvel (Dubois et al. 2018) on the “post-genomic revolution”. With those authors, this chapter shares in particular three fundamental ideas: (i)

THE EXPOSOME AND THE SOCIAL SCIENCES: THE CASE …

255

today, the social sciences cannot do without a reflection on their collaboration or their possible intersections with the life sciences; (ii) thinking the conditions of possibility of interdisciplinary research “[stands] at equal distance from both the sociology of suspicion usually practiced in France and a naively apologetic posture towards any form of biosocial hybridization”; (iii) thinking this way renews the foundational interrogations of the sociological discipline on the boundary between the natural and the social (Dubois et al. 2018, pp. 71–73). The post-genomic era is a time for the social sciences to question about themselves, about what constitutes “the social” and about what can be covered by “the environment”. Following those large epistemological framings, implementing effective tools of inquiry, in specific research fieldworks such as the aetiology and nosology of systemic diseases, inseparably from social justice issues, may be the most fruitful way to deepen the reflexivity of social and biomedical sciences on their possible mutual and own contributions. This chapter is an attempt to do so. Other avenues for future research would also benefit fruitfully from direct and precise discussions between specialists interested in the exposome, at least from three research horizons: biomedical sciences already involved in the promotion of the exposome research enterprise, philosophy and social sciences. This book is a nice step to open this talk.

References Brown, P. (1997). Popular epidemiology revisited. Current Sociology, 45(3), 137– 156. Brown, P., Morello-Frosch, R., Zavestoski, S., & and the Contested Illnesses Research Group. (2012). Contested Illnesses. Citizens, Science, and Health Social Movements. Berkeley, Los Angeles, London: University of California Press. Cavalin, C., Catinon, M., Macchi, O., Vincent, M., Rosental, P.-A., et al. (2021). Expositions aux particules inorganiques : comment poser la question ? In Duwez Emmanuelle, Mercklé Pierre (dir.). Un panel français. L’Étude longitudinale par internet pour les sciences sociales (ELIPSS) (pp. 185–212). Paris: Éditions de l’INED. Cavalin, C., Henry, E., Jouzel, J.-N., & Pélisse, J. (dir.). (2021). Cent ans de sous-reconnaissance des maladies professionnelles. Paris: Presses des Mines. Cavalin, C., Lescoat, A., Sigaux, J., Macchi, O., Ballerie, A., Catinon, M., et al. (2022). Crystalline silica exposure in patients with rheumatoid arthritis

256

C. CAVALIN

and systemic sclerosis: a nationwide cross-sectional survey. Rheumatology, (keac675). https://doi.org/10.1093/rheumatology/keac675 Cooper, G. S., & Stroehla, B. C. (2003). The epidemiology of autoimmune diseases. Autoimmunity Reviews, (2), 119–125. Dubois, M., Guaspare, C., & Louvel, S. (2018). De la génétique à l’épigénétique : une révolution post-génomique à l’usage des sociologues. Revue française de sociologie, 59(1), 71–98. Fassin, D. (2003). Naissance de la santé publique. Deux descriptions de saturnisme infantile à Paris (1987-1989). Genèses, (53), 139–153. Fleck, L. (1979) [1935]. Genesis and Development of a Scientific Fact. Chicago: The University of Chicago Press. Ghiara, V., & Russo, F. (2019). Reconstructing the mixed mechanisms of health: the role of bio- and sociomarkers. Longitudinal and Life Course Studies, 10(1), 7–25. https://doi.org/10.1332/175795919X15468755933353 Giroux, E. (2021a). L’exposome : entre globalité et précision. Bulletin d’histoire et d’épistémologie des sciences de la vie, 28(2), 119–148. Giroux, E. (2021b). L’exposome : vers une science intégrative des expositions Lato Sensu. Revue de la société de philosphie des sciences, 8(3), 9–28. Giroux, E., Fayet, Y., & Serviant-Fine, T. (2021). L’exposome. Tensions entre holisme et réductionnisme. Médecine/Sciences, 37 , 774–778. Joly, P.-B. (2010). On the economics of techno-scientific promises. In Akrich Madeleine, Barthe Yannick, Muniesa Fabian, Mustar Philippe (dir.). Débordements. Mélanges offerts à Michel Callon (pp. 203-221). Paris: Presses des Mines. Juarez, P. D., Matthews-Juarez, P., Hood, D. B., Wansoo, I., Levine, R. S., Kilbourne, B. J., et al. (2014). The public health exposome: A populationbased, exposure science approach to health disparities research. International Journal of Environmental Research and Public Health, 11(12), 12866–12895. Koo, D., & Thacker, S. B. (2010). In snow’s footsteps: Commentary on shoeleather and applied epidemiology. American Journal of Epidemiology, 172(6), 737–739. https://doi.org/10.1093/aje/kwq252 Lescoat, A., Murphy, S., Chen, Y., Vann, N., Galdo, Cella, D., et al. (2022). Symptom experience of limited cutaneous systemic sclerosis from the Patients’ perspective: A qualitative study. Seminars in arthritis and rheumatism, 52. https://doi.org/10.1016/j.semarthrit.2021.11.003 Lescoat, A., Roofeh, D., Townsend, W., Hughes, M., Sandler, R., Zimmermann, F., et al. (2021). Domains and outcome measures for the assessment of limited cutaneous systemic sclerosis: a scoping review protocol. BMJ open, 11(3). https://doi.org/10.1136/bmjopen-2020-044765 Lescoat, A., Cavalin, C., Ehrlich, R., Cazalets, C., Ballerie, A., Belhomme, N., et al. (2019). The nosology of systemic sclerosis: how lessons from the past offer new challenges in reframing an idiopathic rheumatological disorder. The

THE EXPOSOME AND THE SOCIAL SCIENCES: THE CASE …

257

Lancet Rheumatology, 1(4), e257–e264. https://doi.org/10.1016/S26659913(19)30038-4 / https://www.thelancet.com/pdfs/journals/lanrhe/PII S2665-9913(19)30038-4.pdf Lescoat, A., Murphy, S. L., Roofeh, D., Pauling, J. D., Hughes, M., Sandler, R., et al. (2020). Considerations for a combined index for limited cutaneous systemic sclerosis to support drug development and improve outcomes. Journal of Scleroderma and Related Disorders. https://doi.org/10.1177/239 7198320961967 Moore, D. F., Kramer, E., Eltaraboulsi, R., & Steen, V. D. (2019). Increased morbidity and mortality of scleroderma in African Americans compared to Non-African Americans. Athritis Care & Research (Hoboken), 71, 1154–1163. Mukherjee, S. (2016). The Gene: An Intimate History. New York: Simon and Schuster. Murphy, M. (2006). Sick Building Syndrome and the Problem of Uncertainty. Environmental Politics, Technoscience, and Women Workers. Durham and London: Duke University Press. Nathan, N., Montagne, M.-E., Macchi, O., Rosental, P.-A., Chauveau, S., Jeny, F., et al. (2022). Exposure to inorganic particles in paediatric sarcoidosis: the PEDIASARC study. Thorax, 77 (4), 404-407. Epub 2021 Oct. 21. https:// doi.org/10.1136/thoraxjnl-2021-217870 Newman, K. L., & Newman, L. S. (2012). Occupational causes of sarcoidosis. Current Opinion in Allergy and Clinical Immunology, 12(2), 145–150. Oliver, L. C., & Zarnke, A. (2021). Sarcoidosis. An occupational disease? Chest, 160(4), 1360–1367. Parascandola, M. (2011). The epidemiologic transition and changing concepts of causation and causal inference. Revue d’histoire des sciences, 64(2), 243–262. Passeron, J.-C., & Revel, J. (2005). Penser par cas. Raisonner à partir de singularités. In Passeron Jean-Claude, Revel Jacques (dir.). Penser par cas (pp. 9–44). Paris: Éditions de l’EHESS. Proctor, R. N., & Schiebinger, L. L. (2008). The making and unmaking of ignorance. Stanford: Stanford University Press. Rappaport, S. (2011). Implications of the exposome for exposure science. Journal of Exposure Science and Environmental Epidemiology, 21(1), 5–9. Rose, G. (1985). Sick individuals and sick populations. International Journal of Epidemiology, 14(1), 32–38. Rosenberg, C. E., & Golden, J. (eds.). (1992). Framing Disease. Studies in Cultural History. New Brunswick, NJ: Rutgers University Press. Rosental, P.-A. (ed. ), Rosner, D., & Blanc, P. D. (eds. ). (2015). Special issue: From silicosis to silica hazards: An experiment in medicine, history and the social sciences. American Journal of Industrial Medicine, 58(S1), 1–71.

258

C. CAVALIN

Senier, L., Brown, P., Shostak, S., & Hanna, B. (2017). The socio-exposome: advancing exposure science and environmental justice in a postgenomic era. Environmental Sociology, 3(2), 107–121. Susser, M. W., & Susser, E. (1996). Choosing a future for epidemiology: II. From black box to Chinese boxes and eco-epidemiology. American Journal of Public Health, 86(5), 674–677. The EU Framework Programme for Research and Innovation, European Commission, Christopher P. Wild, Miller, G. W., Vermeulen, R., Dillner, J., et al. (2020, 11 February). Launch of the European Human Exposome Network. Understanding the health impacts of a lifetime of environmental https://ec.europa.eu/info/sites/default/files/research_and_inn exposures. ovation/events/presentations/ec_rtd_exposome-event-2020-presentations. pdf Valeyre, D., Prasse, A., Nunes, H., Uzunhan, Y., Brillet, P.-Y., & MüllerQuernheim, J. (2014). Sarcoidosis. The Lancet, 383(9923), 1155–1167. doi: 10.1016/S0140-6736(13)60680-7. Epub 2013 Oct 1. Vineis, P., Avendano-Pabon, M., Barros, H., Chadeau-Hyam, M., Costa, G., Dijmarescu, M., et al. (2017). The biology of inequalities in health: the LIFEPATH project. Longitudinal and Life Course Studies, 8(4), 417–449. https://doi.org/10.14301/llcs.v8i4.448 Vineis, P., & Barouki, R. (2022). The exposome as the science of social-tobiological transitions. Environmental International, 165, 1–8. https://doi. org/10.1016/j.envint.2022.107312. Epub 2022 May 21. Vineis, P., & Kelly-Irving, M. (2019). Biography and biological capital. European Journal of Epidemiology, 34(10), 979–982. Wild, C. P. (2012). The exposome: from concept to utility. International Journal of Epidemiology, 41(1), 24–32. https://doi.org/10.1093/ije/dyr236. Epub 2012 Jan 31.

The Exposome Research Program and Nutrition: The Example of Celiac Disease Paolo Vineis

1

and Antonio Francavilla

The Exposome Concept

The exposome concept is a research program, i.e. a heuristic that is expected to address the relationships between living organisms, their environments and their diseases better than traditional methods. However, the exposome concept still needs development and validation, being born mainly to overcome the limitations of epidemiology, environmental sciences and toxicology. Like other buzz words (for example “holistic” or “complexity”), the exposome concept shows a direction but also includes assumptions. The direction is towards appreciating the interconnectedness of multiple aspects of biography (life-course trajectories, including multiple exposures and socio-economic position) and biology

P. Vineis (B) MRC Centre for Environment and Disease, Imperial College, London, England e-mail: [email protected] A. Francavilla Italian Institute for Genomic Medicine (IIGM), Candiolo, Italy

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 É. Giroux et al. (eds.), Integrative Approaches in Environmental Health and Exposome Research, https://doi.org/10.1007/978-3-031-28432-8_10

259

260

P. VINEIS AND A. FRANCAVILLA

(including underlying molecular mechanisms). One of the main assumptions is that diseases do not depend on single necessary and sufficient causes, but rather on causal networks that start at conception and encompass critical periods in life. Childhood is one of such periods. The exposome focuses on multiple agents (physical, chemical, biological, social) that interact with each other in the whole life course, as opposed to traditional epidemiology that focused on single agents investigated in adults or elderly people (Wild, 2012). From an epistemological point of view, the risk of research programs is that they may fail because they are too abstract, too general and do not allow verification. We can look at the exposome as a double-edged concept: as a research program, it shows the way studies can be designed to avoid a simplified view of disease causation, in particular by considering networks, life-course experiences and underlying molecular mechanisms. As a scientific model of disease, it may turn out to be adequate for specific diseases but inadequate for others. For example, the assumption of the life-course relevance of multiple exposures is not true for highly penetrant gene mutations that cause diseases such as Huntington’s disease, though this case seems to be the exception rather than the norm. The assumption of life-course relevance of multiple exposures is a strong one, since it implies cell memory (likely to be epigenetic) that allows transmission of changes across generations of cells (Vineis 2017). According to Christopher Wild’s original description, there are three components in the exposome concept: specific external exposures (i.e. air pollution), general external exposures (including social interactions and inequalities), and the internal exposome (Wild, 2012). There is a hierarchy among these components, both because the internal exposome is usually (but not necessarily) affected by external exposures, and because the social exposome is an overarching factor that influences all other layers. The internal exposome includes for example metabolic or molecular changes that lead to disease, that can be due to external causes such as dietary exposures, or to internal mechanisms like the generation of reactive oxygen species that occurs as a by-product of lipid peroxidation. Social circumstances act by inducing themselves molecular changes (such as the production of glucocorticoids as a consequence of stress) or because they influence exposures (e.g. occupational chemicals) and behaviors (propensity to smoking or eating unhealthy foods).

THE EXPOSOME RESEARCH PROGRAM AND NUTRITION: …

261

2 Exposome as a Tool to Study the Increased Onset of Autoimmune Diseases If we consider immune dysfunction, one of the main mechanisms behind disease onset, the so-called social-to-biological transition is likely to affect the immune system from early life, including the ways in which it is primed and it functions across the life course, leading to socially patterned immune-related health outcomes (Gares 2017). Let us consider autoimmune diseases (AD) that represent one of the main causes of hospital treatment in Europe. Epidemiological data provide evidence of a steady rise in AD throughout Western societies over the last decades. In a systematic review, 30 studies from the last decades were identified to investigate trends in incidence. Rheumatic, endocrinological, gastrointestinal and neurological autoimmune diseases revealed the following annual % increases per year: 7.1, 6.3, 6.2 and 3.7, respectively (Lerner 2015). Little is known about the molecular basis and epidemiology of autoimmune diseases (AD), and even their classification is very complex (Table 1). Genome-wide association studies have revealed the polygenic basis of AD plus some highly penetrant rare variants (Ye 2018). In fact, most AD are likely to origin from interactions between low-penetrant genetic susceptibility and still largely unknown environmental factors (Costenbader 2012). Let us consider celiac disease (CD), a complex autoimmune disease occurring in about 1% of the population worldwide, with a ratio between women and men varying across countries from 2:1 to 4:1. In CD subjects, gluten ingestion triggers an autoimmune reaction leading to duodenal damage characterized by villous atrophy, infiltration of inflammatory cells and other lesions. CD diagnosis in adults is based on serum IgA antibodies against tissue-transglutaminase 2 (tTG) and confirmed by histopathological analyses of duodenal biopsies. The only efficient treatment for CD is a lifelong gluten-free diet (Lebwohl 2018). Can an exposome approach help in understanding the etiology of CD? The answer is affirmative: the current interpretation of CD is limited to pathogenesis, i.e. the molecular changes that underlie the disease, in particular the cross-reaction between gluten and surface proteins in the duodenum as an effect of autoimmunity. This tells us nothing about etiology, i.e. the causes that trigger autoimmunity. An exposome research could start for example from the observation that the distribution of CD by socio-economic position (SEP) seems to be opposite to what is

262

P. VINEIS AND A. FRANCAVILLA

Table 1 Systemic diseases

Classification of autoimmune diseases Local diseases

Systemic lupus Dermatological erythematosus Sjögren’s Goodpasture’s Syndrome syndrome Scleroderma Sarcoidosis DermatoScleroderma myositis Rheumatoid Psoriasis arthritis Vitiligo Dermatomyositis Alopecia areata

Endocrinologic Neurologic Hematologic Type 1 Multiple Polyarteritis diabetes sclerosis nodosa Myasthenia Idiopathic mellitus Autoimmune gravis thrombocypancreatitis topenic Hashimoto’s purpura thyroiditis Hemolytic Addison’s anemia Antiphospholipid disease antibody syndrome Pernicious anemia

Gastrointestinal Celiac disease Inflammatory bowel disease Autoimmune hepatitis Primary biliary cirrhosis

Note Brief list of major systemic (left) and local (right) autoimmune disease

found for almost all diseases: while the latter—almost with no exception—occur more frequently in disadvantaged SEP, CD seems to occur mainly in higher SEP individuals. In particular, Whyte et al. showed that CD was more common in children living in areas of the UK with higher socio-economic status than in areas with low status (Whyte 2014). High socio-economic status was also positively associated with strong TG2A positivity (odds ratio (OR) 6.85, 95% confidence interval (CI) 1.62 to 28.8), whereas cytomegalovirus (CMV) markers were inversely related to TG2A positivity (OR 0.32, 95% CI 0.12 to 0.84) (Jansen 2017). The latter observation is intriguing, since CMV has emerged recently as a very important player in the modulation of the immune system. Similarly, rotavirus infection has been proposed as viral trigger for several autoimmune processes (Gomez-Rial 2020) including CD onset. A large study found a 33% increased risk of CD autoimmunity within the following 3 months in children who experienced a gastrointestinal infection (hazard ratio (HR), 1.33; 95% CI 1.11 to 1.59). On the contrary, the risk of CD autoimmunity was reduced in children vaccinated against rotavirus (and introduced to gluten before age 6 months) (HR, 0.57; 95% CI 0.36 to 0.88) (Kemppainen 2017). Overall, the evidence on CD and SEP is not entirely consistent and is open to different interpretations. For example, higher CD rates occurred in low-income populations characterized by

THE EXPOSOME RESEARCH PROGRAM AND NUTRITION: …

263

high gluten intake (Catassi 2001, Makharia 2011). Also, the association with SEP was not confirmed in one study in Sweden (Norstrom 2021). These findings highlight the importance of non-heritable factors such as viral infections—in addition to gluten—in shaping an individual’s immune system. But why do diverse phenomena like CD and viral immune modulation differ by SEP? Multiple hypotheses can be thought of, from the simplest one based on selection (richer individuals from high SEP subgroups have a greater likelihood of being diagnosed, particularly for pauci-symptomatic or completely asymptomatic variants), to others based on early life events. In fact, socio-economic status is related to several underlying exposures, such as mode of delivery, early infections and gut microflora. Viral infections (rotavirus) are more likely to occur within lower SEP strata, though this seems not true for CMV (Jansen 2017, Dowd 2012). The interaction between different factors—including gluten intake and infections, according to SEP—and the underlying mechanisms in the onset of diseases is what an exposome approach aims to investigate; evidence reported in this section is an example of gaps in knowledge that the exposome approach could fill.

3 Exposome as a Tool to Study the Gut Microbiome Composition Microbiota play an extremely important role in maintaining human health, and diet is the main factor to regulate the composition and function of gut microbiota. Recent studies have shown that gluten metabolism is closely related to gastrointestinal tract microbiota (Wu 2021). It is intriguing that CD has followed an opposite trajectory compared with acute appendicitis: the latter has steadily declined in the course of time and is likely to be related to the gut microflora and the related immune changes, but follows an opposite association with SEP (more frequent in low SEP). In a geographic study, higher socio-economic status was strongly associated with a lower incidence of uncomplicated appendicitis. This was against the initial study hypothesis, because the authors expected greater access to early diagnosis in areas of higher socio-economic status (Salminen 2020). The gut microflora is a natural suspect for several gastroenteric diseases. We know that its composition is related to early life determinants, but its influence after its stabilization at later childhood is unknown. In the

264

P. VINEIS AND A. FRANCAVILLA

LISA birth cohort, the bacterial profiles were investigated in fecal samples from subjects at age six for the association of microflora with multiple determinants (maternal smoking during pregnancy, mode of delivery, breastfeeding, antibiotic treatment between one and two years of age, gender and socio-economic status). The strongest associations were with feeding practice and antibiotics use (Gschwendtner 2019). SEP was not associated with microbiome status. The interpretation of the authors is that the analysis of the gut microbiome composition after its stabilization indicates the persistence of early life determinant effects up to six years, although the effects were not pronounced and are probably overridden by other lifestyle and environmental factors becoming more important with increasing age. In another study in the United States, the gut microbiome of neonates and infants was investigated. Multiple factors were associated with neonatal composition including maternal race-ethnicity, breastfeeding, mode of delivery, marital status, exposure to environmental tobacco smoke and indoor pets. These findings were replicated in the infants, with similar patterns (Levin 2016). In a review, Harrison and Taren (2018) discuss the role of SEP in modulating the health disparities seen between lower-income and higher-income populations, via a different composition of gut microflora (Harrison 2018). As described in Fig. 1, they hypothesize that multiple components of the external exposome influence the composition of the gut microflora that in turn leads to chronic inflammation and the risk of non-communicable diseases. Figure 1 is a typical example of a research program inspired by the exposome concept. Multiple exposures are considered, under the overarching influence of SEP, and can be explored with “internal exposome” tools, such as metabolomics, metagenomics and epigenetics. Unfortunately, a scientific approach requires more than the coordinated use of multiple tools exploring the external and internal exposome in large cohorts, as is done now in consortia. It also requires the possibility of falsifying hypotheses and disentangling alternative hypotheses. Here come the difficulties. It seems, in fact, that the very same schemes generically proposed for non-communicable diseases (NCD) (like in Fig. 1) are valid for childhood obesity and even autoimmune diseases, i.e. the schemes are too broad and cannot be fine-tuned to explain specific conditions or diseases. For example, what triggers the immune response against the duodenal mucosa after its exposure to gluten is still unexplained,

THE EXPOSOME RESEARCH PROGRAM AND NUTRITION: …

265

Fig. 1 In high-income countries (HIC), individuals with low socio-economic status (SES) are presented with several significant complications that may both directly and indirectly influence the composition of the microbiota and, concurrently, the immune system. Specifically, we suggest that environmental stressors, the pressure to eat affordable food that is filling and palatable, along with alterations in health care and medication use, can hamper microbiota diversity and promote a low-grade inflammatory state that precipitates metabolic disease. NCD, non-communicable disease; T2D, type 2 diabetes (reproduced from Harrison and Taren 2018)

though a hypothesis has been put forward that the sequence from breastfeeding to the intake of gluten-based food may be crucial. This is a specific hypothesis that can be falsified with good exposome studies. The fact that apparently CD follows an opposite trend in its distribution by SEP, compared to almost all other diseases, is a precious observation because it can contribute to narrowing down the existing hypotheses on CD.

4

Conclusions

In summary, the exposome is a powerful concept but it needs more specificity. After the era of “agnostic” investigations, in which we have developed broad theories of etiology and pathogenesis, we now need a new generation of exposome studies that are more hypothesis-driven and have greater granularity, possibly taking stock of peculiar observations such as the association between SEP and CD (if confirmed). This

266

P. VINEIS AND A. FRANCAVILLA

type of proposed approach is not a way of returning to earlier methodologies and paradigms which analyzed risk factors one by one rather than their multiple associations. In fact, the exposome approach and an analytical approach are complementary and interact in a virtuous circle. The ability of the exposome research to generate new hypotheses can be ideally followed by the analytical comparison of the hypotheses, then followed again by a more holistic integration of different causal and mechanistic pathways. Acknowledgements Paolo Vineis is a member of the Health Protection Research Unit in Chemical and Radiation Threats and Hazards, a partnership between Public Health England and Imperial College London which is funded by the National Institute for Health Research (NIHR). We are grateful to Barbara Pardini and Alessio Naccarati for critical reading of the chapter.

References Aaron Lerner, P. J., Torsten Matthias. The world incidence and prevalence of autoimmune diseases is increasing. International Journal of Celiac Disease 3, 151–155, https://doi.org/10.12691/ijcd-3-4-8 (2015). Catassi, C., Doloretta Macis, M., Ratsch, I. M., De Virgiliis, S. & Cucca, F. The distribution of DQ genes in the Saharawi population provides only a partial explanation for the high celiac disease prevalence. Tissue Antigens 58, 402–406, https://doi.org/10.1034/j.1399-0039.2001.580609.x (2001). Costenbader, K. H., Gay, S., Alarcon-Riquelme, M. E., Iaccarino, L. & Doria, A. Genes, epigenetic regulation and environmental factors: which is the most relevant in developing autoimmune diseases? Autoimmunity Reviews 11, 604– 609, https://doi.org/10.1016/j.autrev.2011.10.022 (2012). Dowd, J. B., Palermo, T. M. & Aiello, A. E. Family poverty is associated with cytomegalovirus antibody titers in U.S. children. Health Psychology 31, 5–10, https://doi.org/10.1037/a0025337 (2012). Gares, V., Panico, L., Castagne, R., Delpierre, C. & Kelly-Irving, M. The role of the early social environment on Epstein Barr virus infection: a prospective observational design using the Millennium Cohort Study. Epidemiology and Infection 145, 3405–3412, https://doi.org/10.1017/S0950268817002515 (2017). Gomez-Rial, J., Rivero-Calle, I., Salas, A. & Martinon-Torres, F. Rotavirus and autoimmunity. Journal of Infection 81, 183–189, https://doi.org/10.1016/ j.jinf.2020.04.041 (2020).

THE EXPOSOME RESEARCH PROGRAM AND NUTRITION: …

267

Gschwendtner, S. et al. Early life determinants induce sustainable changes in the gut microbiome of six-year-old children. Scientific Reports 9, 12675, https:// doi.org/10.1038/s41598-019-49160-7 (2019). Harrison, C. A. & Taren, D. How poverty affects diet to shape the microbiota and chronic disease. Nature Reviews Immunology 18, 279–287, https://doi. org/10.1038/nri.2017.121 (2018). Jansen, M. A. E. et al. Ethnic differences in coeliac disease autoimmunity in childhood: the Generation R Study. Archives of Disease in Childhood 102, 529–534, https://doi.org/10.1136/archdischild-2016-311343 (2017). Kemppainen, K. M. et al. Factors that increase risk of celiac disease autoimmunity after a gastrointestinal infection in early life. Clinical Gastroenterology and Hepatology 15, 694–702 e695, https://doi.org/10.1016/j.cgh.2016.10.033 (2017). Lebwohl, B., Sanders, D. S. & Green, P. H. R. Coeliac disease. Lancet 391, 70–81, https://doi.org/10.1016/S0140-6736(17)31796-8 (2018). Levin, A. M. et al. Joint effects of pregnancy, sociocultural, and environmental factors on early life gut microbiome structure and diversity. Scientific Reports 6, 31775, https://doi.org/10.1038/srep31775 (2016). Makharia, G. K. et al. Prevalence of celiac disease in the northern part of India: a community based study. Journal of Gastroenterology and Hepatology 26, 894– 900, https://doi.org/10.1111/j.1440-1746.2010.06606.x (2011). Norstrom, F. et al. Family socio-economic status and childhood coeliac disease seem to be unrelated-A cross-sectional screening study. Acta Paediatra 110, 1346–1352, :https://doi.org/10.1111/apa.15562 (2021). Salminen, P. Acute Appendicitis incidence-predisposing factors, from microbiota to socioeconomic status? JAMA Surgery 155, 338–339, https://doi.org/10. 1001/jamasurg.2019.6031 (2020). Vineis, P. et al. Epigenetic memory in response to environmental stressors. FASEB Journal 31, 2241–2251, https://doi.org/10.1096/fj.201601059RR (2017). Whyte, L. A., Kotecha, S., Watkins, W. J. & Jenkins, H. R. Coeliac disease is more common in children with high socio-economic status. Acta Paediatra 103, 289–294, https://doi.org/10.1111/apa.12494 (2014). Wild, C. P. The exposome: from concept to utility. International Journal of Epidemiology 41, 24–32, https://doi.org/10.1093/ije/dyr236 (2012). Wu, X., Qian, L., Liu, K., Wu, J. & Shan, Z. Gastrointestinal microbiome and gluten in celiac disease. Annals of Medicine 53, 1797–1805, https://doi.org/ 10.1080/07853890.2021.1990392 (2021). Ye, J., Gillespie, K. M. & Rodriguez, S. Unravelling the roles of susceptibility loci for autoimmune diseases in the Post-GWAS Era. Genes (Basel) 9, https:// doi.org/10.3390/genes9080377 (2018).

Index

A autoimmune diseases (AD), 191, 247, 250, 261, 264

B biologization, 4, 66, 67, 69, 71, 88, 89, 91, 92, 146, 230 biomarker, 20, 36, 37, 39, 43, 53, 66, 87, 109, 131, 138, 140, 142–144, 147, 149, 151, 152, 156–159, 162, 177, 184, 185, 226, 227, 229, 230 biosocial, 3, 4, 11–19, 21, 23–26, 133 Bourdieusian, 245

C cancer, 35, 37, 39, 42, 45, 55, 133–135, 173, 190–192, 221, 252 causation, 39–41, 46, 48, 49, 52, 56, 130, 132, 153, 155, 156, 159, 163, 182, 190, 195, 251, 260

D data integration, 102–105, 108–110, 112, 113, 115–120, 230 disease non-communicable disease (NCD), 176, 179, 180, 264, 265 E eco-epidemiology, 194, 195 eco-social theory, 148, 160 embodiment, 3, 4, 20, 49, 83, 85, 148–151, 159, 160, 162, 245, 246 environmental health, 129, 130, 132, 146, 173, 175, 210, 230, 253 epigenetics, 9–26, 66, 137, 144, 145, 150, 174, 260, 264 European Human Exposome Network (EHEN), 154, 228, 241 evidence, 3, 4, 12, 14, 17, 18, 20, 23, 25, 35, 39, 40, 54, 56, 65, 66, 75, 92, 103, 110, 114, 116, 118, 121, 134, 137, 150, 155–158, 218, 231, 251, 261–263

© The Editor(s) (if applicable) and The Author(s) 2023 É. Giroux et al. (eds.), Integrative Approaches in Environmental Health and Exposome Research, https://doi.org/10.1007/978-3-031-28432-8

269

270

INDEX

exposome external exposome, 142, 146, 149, 162, 228, 264 internal exposome, 137–140, 142, 144, 149, 162, 163, 176, 184, 260, 264 exposomics, 66, 131–133, 141–145, 147, 149–151, 154–158, 173, 174, 177–179, 181, 183–185, 189, 194, 200, 226–231, 241–243, 248, 252, 254

G genome, 10, 11, 14, 15, 106, 130, 135, 141, 174–176, 179, 180, 184, 185, 188, 193, 199, 239, 240 genome-wide association studies (GWAS), 141, 180, 240, 261 global health, 102, 105, 115–120 gut microbiome, 264

H health geography, 210, 213, 214, 226, 229, 232 health inequalities, 40, 66, 67, 81, 82, 89, 90, 92, 159, 161, 163, 215, 221, 223, 224, 243, 244 holism, 130, 146, 241 holistic, 14, 17, 130–132, 140, 142, 144–147, 161–163, 181, 210, 216, 219, 221, 223, 228, 229, 231, 241, 245, 259, 266 Human Exposome Project, 141, 179 Human Genome Project, 11, 140, 179, 240

I incommensurability, 4, 12, 15, 18, 19, 25, 26

individual, 10, 17, 20, 21, 23, 24, 36, 38–48, 51–56, 69, 73, 76, 80, 82, 85, 88, 90, 91, 102, 104, 106, 107, 109, 112, 113, 115, 116, 120, 129, 133, 134, 136, 137, 142–148, 152, 153, 156, 159–163, 175, 176, 180, 181, 184, 187–190, 194, 195, 197, 199, 214–218, 222, 225, 226, 229, 230, 240, 244, 245, 247, 248, 262, 263, 265 information transmission, 46, 50, 51, 53 integration, 3, 19, 102–108, 111, 113, 115, 117–119, 130, 132, 133, 139, 145, 147–152, 154, 156, 158, 159, 161–163, 181, 182, 226, 227, 241, 266 integrative approaches, 11, 18, 57, 147, 148, 150, 153, 159, 162, 214, 223, 226, 228, 229, 231 interdisciplinarity, 114, 139, 145, 147, 162, 253 interdisciplinary, 11, 12, 17–19, 130, 147, 149, 161, 240, 242, 253, 255

M molecular epidemiology, 35, 36, 39, 41, 45, 49, 56, 131, 139, 141, 144, 145, 147, 149–151, 158, 161, 162 moral economy, 87, 88, 91, 92

N niche, 185, 187–189, 192, 199 niche construction, 187, 190 obesogenic niche, 175, 199 pathogenic niche, 175, 190, 192–194, 196, 199, 200

INDEX

O obesity, 35, 40, 173, 182, 196–198, 218, 219, 264 P philosophy of biology, 4, 11, 14, 175, 185, 188, 189, 196, 199 planetary Health, 102, 105, 111–116, 119–121 pluralism, 19, 132, 134, 147, 157, 242, 245 population, 2, 20, 41, 46–48, 54, 55, 66, 69, 72, 73, 76–81, 84, 85, 91, 102–106, 109–120, 134, 135, 144, 145, 147, 152, 153, 160, 161, 163, 173–176, 180, 182, 185, 186, 188–200, 211, 213–215, 220, 222–225, 228, 230, 231, 240, 244–247, 249, 250, 261, 262, 264 postgenomics, 102, 104 process, 10, 11, 13–21, 23–26, 36–42, 44–53, 55, 56, 84, 86, 90, 106, 107, 110, 137, 142, 147–149, 152, 153, 158, 161, 177, 188, 212, 219, 221, 223, 229, 262 psychosocial, 20, 47, 66, 67, 71, 75–84, 86, 87, 89–91, 160, 181 public health exposome (PHE), 142, 146, 147, 178, 243, 244 R reductionism, 10, 16, 25, 48, 49, 110, 119, 142, 145, 148, 161, 185, 231, 241

271

S salutogenesis, 46–48, 53, 55 Science and Technology Studies (STS), 11, 13–15, 18, 20, 24–26, 244 social determinants of health, 14, 17, 20, 69, 83, 105, 112, 116, 145, 150, 242, 244 social epidemiology, 4, 26, 40, 65–67, 75, 77–79, 83, 85, 87, 89, 91, 131, 145, 147, 149–152, 158, 160, 161, 163, 195, 215, 248, 249 socio-exposome, 146, 147, 243, 244 socio-marker, 38, 39, 41, 43–48, 50–57 spatial data, 223, 226, 228, 230, 231, 243 spatial epidemiology, 111, 210 spatial inequalities in health, 214, 219–224 stress, 20, 22, 38, 66, 67, 69, 73, 75, 76, 78, 81–84, 86, 87, 89, 90, 106, 110, 136, 159, 160, 177, 181, 196, 260 style(s) of reasoning, 12, 18, 23, 25, 26 systemic disease, 242, 246–249, 251, 252, 254, 255

T toxicology, 106, 109, 131, 133–135, 162, 173, 189, 241, 245, 259